This is part 20 of 30 of Hammertime. Click here for the intro.
There’s a serious and scary phenomenon which Valentine’s recent posts have been touching on: much of who you are only exists (or is expressed) in the presence of other people. In the words of Bishop Berkeley, esse est percipi: To be is to be perceived. Hammertime will always be an incomplete endeavor unless it is applied to social settings – there are major chunks of the psyche only accessible in such settings.
Up to now, Hammertime has mostly been a set of tools for the individual rationalist in a social vacuum. Today I want to talk the problem of other human beings, and how to go about designing social interactions that are conducive to the practice of instrumental rationality.
Hammertime Day 20: Friendship
Background: The Intelligent Social Web
There’s good evidence in biology that the power of the human brain largely evolved to solve ever-complexifying social problems. Much of the heavy cognitive machinery in your head is primarily built for and responds best to social interaction. Brains are extremely good at detecting social threats and anomalies, at regulating implicit status ladders, at reading body language, and at simulating other brains.
This post is a start at the Design of optimal two-person interactions.
Iterated Games
Rationalists spend a lot of time railing against the failings of causal decision theory, and promoting alternatives that solve them. The uncomfortable truth, however, is that you will not make causal decision theorists cooperate on the prisoner’s dilemma by throwing tomes of philosophy at them, and many many people are causal decision theorists. Not all hope is lost though: there’s a known, albeit unglamorous, solution to coordination failures within the framework of causal decision theory: iterated games.
Iteration is the easiest path to building strong friendship: make interactions longer and more regular.
In the middle of January, I began contacting friends and setting up regularly weekly chats. Almost nobody refused. A handful of interactions fizzled out, but the ones that lasted have been unbelievably positive. I kept ramping up the number of conversations until it felt actively fatiguing. Today this habit alone allows me to talk to an average of one extra person per day for an hour and a half.
Human beings are unbelievably reciprocal creatures in stable long-term relationships. The incentives are quite robust. Jordan Peterson once highlighted this with a pithy phrase about marriage (paraphrased): “You can’t win an argument against your wife if she loses. After all, you still have to live with her.”
Of course, human beings are also stupid and perverse enough to ignore the strongest incentives. How many millions of life-long partnerships ended with decades of mutual abuse? Keep your eyes open.
Conversation 101
Here are three object-level ideas for having useful conversations.
Socratic Ducking
Rubber Ducking
Getting a person to act as a rubber duck who you talk your ideas out to in order to get a clear handle on them.
Socratic Ducking
Aiding a partner in thinking through an idea or solving a problem. Combines socratic questioning and rubber ducking. Attempt to offer few suggestions and thoughts while instead alternating between stimulating questions and attentive silence. Encourage the other person to think through complex threads and think deeply about the ramifications of ideas and possible solutions.
Oftentimes there is a clear listener and talker in a conversation. As the listener, focus primarily on paying attentive silence and occasionally asking pointed or clarifying questions when the conversation seems to dry up. The primary goal is to keep your partner generating ideas and on track.
A friend of mine stimulated a major breakthrough in a session of aversion factoring for me by nodding silently the whole time, except for uttering a single well-timed word: “try!” This encouraged me to expend the necessary mental effort to break through that mental barrier and correctly identify an aversion towards planning.
ITTs
The Ideological Turing Test is a concept invented by American economist Bryan Caplan to test whether a political or ideological partisan correctly understands the arguments of his or her intellectual adversaries: the partisan is invited to answer questions or write an essay posing as his opposite number. If neutral judges cannot tell the difference between the partisan’s answers and the answers of the opposite number, the candidate is judged to correctly understand the opposing side.
Intellectual (Ideological?) Turing Tests, or ITT’s, can be rather laborious. The minified conversational norm is: you are not allowed to move forward with an argument until you have accurately summarized the other person’s point of view to their satisfaction.
Rabbitholes
Conversations can get derailed rather rapidly, and it’s a well-established fact that all conversations after midnight will devolve into a debate about consciousness.
For online conversations, I make a habit of collecting possible tangents on a sheet of paper when they come to mind, instead of immediately tossing them into the fray and risking the entire current train of thought. There will always be time later for your fascinating point.
Take a Yoda Timer to train the following TAP: whenever a related conversation topic comes to mind, ask yourself whether you want to go down that rabbit-hole.
Daily Challenge
Book a 30 or 60-minute chat with me on Calendly to talk about anything.
[Update: I am closing the Calendly as of 10/2024, as my availability has become less predictable. Please reach out directly on LW if you would like to chat.]
I read something once—I don’t remember where or what it was called—about how to design good buildings where people can get really good work done, and it heavily emphasized the importance of chance encounters: people need to be able to talk to each other, so they can sow the seeds of collaboration and so forth, in a way that’s as low cost as possible, and also shuffles people around in an interesting way. It’s much smoother for us to start collaborating by running into each other in the hallway and asking each other what we’ve been up to than for me to have to sit down and figure out who I want to talk to, then finding them in person (let alone emailing them to set up a time to meet—what a nightmare).
Later I read another thing that emphasized the importance of repeated chance encounters for naturally building friendships. My model is something like, if you know you’re going to randomly run into this person over and over again, you can’t just be neutral towards them: your social machinery has to decide whether the equilibrium in your iterated PD is going to be mostly cooperation or mostly defection.
Some corollaries: living alone is terrible, working remotely is terrible unless your coworkers are terrible, third places where you can run into people you like or will like randomly are incredibly important.
I don’t know if this is what you read, but this reminds me of Bell Labs:
― New York Times
The thing I read was either about Bell Labs, Google, Pixar, or a lab at MIT. I may have seen multiple articles make this point.
In Subduing Moloch, Teja suggests intentionally creating a channel for rationalists to have one-on-one conversations with each other. As a result, he and I have already had a video chat, and we’ve joined the LessWrong Slack in order to determine if that might be an appropriate venue to build this project.
I intend to book a conversation with you, and I will also consider creating a similar Calendly system for people to book time with me.
When I imagine someone like this, I imagine someone who does not know what to do with themselves when they are alone, however briefly, and finds any length of time alone unpleasant. I imagine someone who scarcely has any sort of self at all, a zombie. Not the shambling dullard of horror movie zombies, but someone full of activity and apparent life yet with no-one at home.
But this is just my imagination confabulating by the yard. I do not actually know what it is like to be such a person. However, the number of modern Buddhists and the like proclaiming their discovery that they have no self (e.g. Sam Harris, Susan Blackmore) suggests to me that there may be something to this picture. And how else can one explain the existence of behaviourists, other than as people with no awareness of their own self, like other people who lack a sense of smell and may never realise they’re missing anything?
For me it is the opposite. Dealing with people is always irksome, however useful it may be. What I am is most fully expressed when alone.
I feel similarly to what you expressed in your first paragraph, and somewhat similar to your third. When I realize certain people can’t stand being alone, I imagine them as someone who has no idea what to do with themselves. I feel like my brain is highjacked if I’m not given enough alone time to process my thoughts, and that parts of me are never fully expressed until I am alone.
Maybe this means I need to improve my social circle?
Your observation of the Buddhist “no self” claims seems to me like a misunderstanding due to different definitions of self. After much staring at these claims on my part, I think what they are (rightly) saying is that there is no single “executive” module in charge in our brains, and that our impression of a unified self is an illusion that rises from a bunch of separate modules.
Is that different from saying that a car does not exist, because it’s made from a bunch of separate modules? And what of the modules, that are made of smaller parts, all the way down to atoms, quarks, and whatever may lie below those? I don’t see a level to stop once you say that a thing does not exist because it is made of parts.
I go with the view that a car does exist, and so do I, even though I am made of parts.
Yes; see the first half of this comment for an explanation. The physical existence of the entities is not in question.
And yet, there is still a claim that there is something delusional about the notion of an “I”. Perhaps my puzzlement with all this sort of thing (enlightenment, kensho, etc.) is not that I do not understand the state of mind being boomed (at least, when described in terms of concrete experiences rather than nonexistence), but that I do not understand the state that is being booed.
One way of putting it, which I think should be mostly accurate, is that the state that is being a booed is a belief in the homunculus fallacy.
Dennett, Kurzban, and others have pointed out that there are facts about the way in which the mind and consciousness function which feel deeply counter-intuitive, and that even neuroscientists and psychologists who in principle know that the brain is just a distributed system of separate modules, still often seem to operate under an intuition that there is a single “central” self (as seen from some of the theories that they propose).
I’m not sure whether that’s the source of the intuition, but it also seems related that humans seem to have a core system for reasoning about agency which takes as an axiom the assumption that agents exhibit independent, goal-directed motion (as opposed to objects, which only act when acted upon). Which makes sense if you’re just reasoning about e.g. social dynamics, but gets you into trouble if you try to understand the functioning of the brain and feel intuitively convinced that there has to be a “central agent” (homunculus) there somewhere, and it can’t just be interacting objects all the way down. It’s been a while since I read it, but IIRC Kurzban’s book had a bunch of examples about how neuroscientists who should know this stuff were still making hypotheses that had the homunculus intuition lurking somewhere.
So when Buddhists say that “there is no self”, they are saying that the intuitive belief in the homunculus is wrong; and when talk about realizing that this is a delusion, they talk about actually coming to internalize this on a deep level.
There are those concepts “just” and “central” again. Is a car “just” a system of separate modules? Is there a “single, central, car”? To me these questions are as otiose as the elementary conundrum about whether a tree falling unheard makes a sound.
The car consists of the right components arranged in the right way. When they are, a mode of operation is created that cannot be found in any part. There is the car, as real as anything. “The car” is not a separate animating principle that makes the components perform. Knowing how the car works does not dissolve the car, but does enable one to use it more effectively, and to repair it when it goes wrong.
So also with the mind.
A complication present with the mind but not the car is that (with present technology) the car is not the person operating it, whereas the mind and its operator are one and the same. We must discover and mend our flaws with the very instrument that is flawed. That is what makes it difficult, but there are ways. Some of which I have experienced, although not along the lines of recent posts about enlightenment experiences.
I haven’t read Kurzban’s book*, but I agree that there’s a lot of dualist talk from supposed materialists, when they talk about how “you” don’t do things, “your brain” does them, and the like. Also the anti-pattern “X is the mechanism of Y, therefore Y does not exist”, X and Y being physical and mental phenomena respectively.
\* (The free extract on Amazon suggests to me that the book will argue for the nonexistence of the mind.)
This was a failure of communication on my part. While indeed there are extroverts for whom what you say about zombieness is true, I’m making a general claim that everyone has a good deal of neural circuitry (you can call it persona or masks if you like) that only expresses itself in the presence of others. Whether or not you want to identify this as “what you are” is up to you. For what it’s worth I think you’ve made the right choice.
Nevertheless, it’s useful to optimize the social circuitry and that’s can only really be done while in social situations. That’s all I’m saying.
This was a great post, and I fully agree with the idea that working to make our interactions with others positive is an important part of making life good. I found that a good additional exercise to go with this post was to set a Yoda timer and reach out to friends who I haven’t talked to in a bit.
Socratic Ducking
I have a habit of “ad-hoc Socratic Ducking” in the sense that I will talk out loud to someone about an idea I’m thinking of and explore it in real time, trying to see what it’s like and figure it out. The problem is that it can seem like I’m stating definite opinions about how the world works or a specific direction to take a project. This can get weird if it’s a taboo topic or a project that involves the person I’m talking to.
I feel like the key with this is awareness and encouragement. Both parties should be aware that they’re doing it, otherwise it can get complicated. If my partner doesn’t encourage me to go deeper, I may get self-conscious and turn off the idea-generator.
ITT
I tried to do an abbreviated ITT once—we ended up arguing about the setup/rules (or doing it at all) instead. Many people seem to not even want to pretend they could understand the other side’s viewpoint, as if it’s so morally repugnant that the ideas are unspeakable. I think this is often a thin veil to cover up for the mismatch between the fragility of one’s stance and the fervor they defend it with.
Rabbitholes
Some of my favorite conversations were with my college roommate as a freshman. Time became a limit—we had to choose what to talk about, otherwise we’d talk through the night. It felt like an all-you-can-eat-buffet of ideas and interesting thoughts—we started talking about something but then we’d see a new shiny thought somewhere else. So we started writing down the branches of our conversations so that we wouldn’t get too lost. I don’t think we ever really revisited it but it was a nice way to feel okay with deferring a certain branch of the conversation.
A lot of those branches still feel very open. Even though we haven’t talked about some of them for years, we’ve been able to pick right back up on a few of them.
Looking forward to talking to you!
Your Calendly page currently only has 60-minute meetings available.
Just book a slot and leave a message for how long you want it to be.