Social status part 2/2: everything else
2.1 Post summary / Table of contents
This is the second of two blog posts where I try to make sense of the whole universe of social-status-related behaviors and phenomena. The previous one was: “Social status part 1/2: negotiations over object-level preferences”.
In that previous post, I was focusing on the simplified special case of two people interacting, where they have different “object-level preferences”, but where we ignore their higher-level preferences regarding the interaction itself (e.g. the desire not to be rude). In that post, I offered a series of vignettes and toy models, covering the ideas of “mostly leading” versus “mostly following”, “ask versus guess culture”, “pushiness”, and some other topics.
Now that we have that foundation, we’re ready to talk about the universe of social behaviors built on top of that basic dynamic, including dominance-versus-prestige, getting offended, passive-aggressiveness, status, self-deprecation, and more.
This post is organized as follows:
Section 2.2 surveys the wide variety of reasons why people might want to “mostly lead” or “mostly follow” in an interaction. I offer six categories. First, leading is nice for the obvious reason that you get to fulfill your object-level desires (see previous post). Second, it’s possible that there is an “innate drive” to lead. Third, you may want to “lead” or “follow” if you think your object-level desires are better- or worse-informed than the other person’s, respectively. Fourth, you may want to “follow” due to general admiration of the other person, or “lead” due to general scorn of the other person. Fifth, you may want to “follow” to avoid blame, or “lead” to get credit. Sixth, you may want to accommodate (or undermine) the other person’s preference to lead or follow.
Section 2.3 introduces “dominance versus prestige” of Dual Strategies Theory, and how they relate to the idea of “mostly leading” versus “mostly following” from the previous post. Specifically, “Alice is being dominant” means: Alice is “mostly leading”, and Beth wishes that Alice wouldn’t lead so much, but Beth doesn’t do anything about it because she’s afraid of offending Alice. By contrast “Alice has prestige” means: Beth directly wants to “follow”, and wants Alice to lead, because Beth likes and admires Alice.
Section 2.4, following this 2018 blog post, defines “status” as a tacit shared understanding about how much a person can lead in a certain context without getting pushback.
Section 2.5 offers some elaborations on dominance-flavored interactions, including an argument against some intuitions associated with the term “pecking order”, what people mean by “taking offense”, why and how people are passive aggressive (and why the targets often find that really annoying), and how people wind up being dominant.
Section 2.6 offers some elaborations on prestige-flavored interactions, including speculations on how it works psychologically, how people wind up being prestigious, the relation to reciprocal altruism, and the related phenomena of competitive generosity and self-deprecation.
Section 2.7 is a brief conclusion.
2.2 Six reasons that someone might want to “mostly lead” or to “mostly follow”
In the previous post, I talked about “mostly leading” and “mostly following” as a higher-level description of a dynamic that occurs in negotiations over conflicting object-level preferences (e.g. whether to order pizza versus sushi).
However, people are also aware (implicitly or explicitly) of the very fact that they are leading versus following in an interaction, and consequently can have preferences about that higher-level social dynamic.
I think there are many reasons that people might want to move in the direction of leading or following. Here are the six categories that I could think of.
2.2.1 The obvious instrumental reason that everyone wants to lead
This one is obvious: “Leading” amounts to getting all your object-level preferences satisfied, whatever they are, and wherever they come from! So other things equal, of course you want that! For example, you have an innate hunger drive, and if you’re “mostly leading” then you get to decide what and how and when to eat. And likewise, you have an innate curiosity drive, and if you’re “mostly leading” then you get to pick whatever topic of conversation best satisfies your curiosity. And so on, for every one of your desires. All these things are (tautologically) motivating.
This constitutes a universal pro tanto reason to want to lead more.
2.2.2 Is there an “innate drive” to “mostly lead”?
Above I was arguing that leading is instrumentally-useful in the context of within-lifetime learning and planning. I was not talking about evolution, i.e. I was taking our innate drives and dispositions as a given.
But there is in fact a separate possibility that we (also) have an innate drive to lead more, as an end in itself, built into our brain by our genome, and that this drive is there because leading more is instrumentally-useful towards the end of inclusive genetic fitness.
Here’s an analogy to explain this distinction: the common human behavior of not-falling-off-cliffs. This behavior comes about via two independent mechanisms:
First, in the context of within-lifetime learning and planning, it’s instrumentally useful not to fall off cliffs. That means: we tend to take precautions around cliffs thanks to reason, planning, and/or experience—we know that falling off a cliff would be painful and bad in lots of ways, so we strive to avoid doing so.
Second, it is also true that, in the context of evolution, it’s instrumentally useful to not fall off cliffs—e.g., if we fall and die, then we won’t have children. Therefore evolution built into our brains an innate fear of heights. When you look directly down off a precipice, you tend to feel some anxiety, and maybe a funny tingling sensation.
These are two different mechanisms, arising from two quite different algorithms running in different parts of the brain. They are both involved in the behavior of not-falling-off-cliffs, but for other behaviors, they come apart: the common behavior of “not playing with loaded guns” is a result of within-lifetime learning and planning, not innate drives; and conversely, the common behavior of “not playing with large harmless spiders” is a result of innate drives, not within-lifetime learning and planning.
So again: insofar as people want to “mostly lead”, are there innate drives involved?
My answer is: I strongly believe (for reasons discussed in §4.4.1 here) that, insofar as people want to “mostly lead”, there are innate drives involved, and not exclusively within-lifetime learning, although the latter is obviously relevant too. But I don’t currently think there’s an innate drive to “mostly lead” per se. Rather, I think there’s an innate drive that we might loosely describe as “a drive to feel liked / admired” (discussed in more detail in §4.4 of my valence series), and also an innate drive that we might loosely describe as “a drive to feel feared”. (Like all innate drives, these are each much stronger in some individuals than others.) These drives are just upstream of gaining an ability to “mostly lead”, for reasons that should be obvious (or if not, keep reading this post). I’m very interested in fleshing out exactly what these innate drives are and how they work in the brain, but it’s outside the scope of this series.
What about the converse? In principle, some people could have an innate aversion to leading, perhaps related to “being shy and tentative in social situations”. My current guess, as above, is that there is no such innate aversion, per se. Instead, I think “leading” can sometimes give rise to consequences that are strongly aversive for other reasons—see §2.2.5-§2.2.6 below for how “leading” can lead to blame, guilt, shame, and pissing people off. Those upstream aversions are of course related to various innate reactions, which are stronger in some people than others. I think shy people find these things sufficiently aversive that they bend over backwards to avoid even a small chance of experiencing them. (That’s a guess based largely on my own experience; please comment if it doesn’t ring true to you.)
You might notice that I’m avoiding a common thing that evolutionary psychologists do (e.g. Secret of Our Success by Henrich), which is to point to particular human behaviors and just say that they’re evolved—for example, they might say there’s an “innate drive to be a leader”, or “innate drive to be dominant”, or “innate drive to imitate successful people”, and so on. I think those are basically all “at the wrong level” to be neuroscientifically plausible (as such), in a sense that I explain in §13.2.2 of this older post. For example, I don’t think there’s an “innate drive to imitate successful people” per se; instead I think there’s a much much more general and ancient brain mechanism discussed in post 2 of my valence series, for which imitating-successful-people is partly a special case (see §4.5 of my valence series); and then I think that very general mechanism is supplemented by some specific low-level adaptations, e.g. an innate brainstem reflex to look at people’s faces (see Johnson 2015). See also §2.6.1 below.
2.2.3 You may want to lead more or less when you think your desires are better- or worse-informed than the other person’s, respectively
Let’s jump right into an example.
Example scenario: {Alice = Expert tour guide, Beth = Novice tourist}: I am Beth, and I’m visiting a dangerous country where I don’t know the culture or speak the language. But I’m with an experienced tour guide Alice. As we walk around, I’ll tend to “mostly follow” in my interactions with Alice, regarding where to go and what to do. Why? Because I trust Alice’s object-level preferences much more than I trust my own. After all, my own ill-informed object-level preferences for where to go and what to do might get me lost or killed, whereas Alice’s object-level preferences are a path to a safe and enjoyable trip.
For example, if I see a river that looks cool, then maybe I have a strong object-level preference to go check it out. I might mention that preference to my tour guide Alice, but I would do so in an extremely tentative, as opposed to pushy, manner. After all, if Alice has an object-level preference to stay away from that river, then probably she has a very good reason for that preference! Maybe the river is in the wrong direction from where we need to go. Heck, maybe the river is full of piranhas! Who knows! So I really want Alice to “mostly lead”, and not defer to my far-less-informed desires.
Notice that my desire for Alice to “mostly lead” here is quite context-dependent. If Alice is guiding me through the country’s confusing subway system, then I definitely want Alice to take the lead. But if Alice starts making dumb racist jokes, I might immediately start leading more myself by pushing back, especially if the tour is over and I no longer need to be in her good graces.
Conversely, if we switch perspectives, then we see that the tour guide Alice wants to “mostly lead”, and wants Beth to “mostly follow”, in regards to decisions about where to go and what to do, by the same logic as above: Alice knows that Beth’s preferences are ill-informed, while her own preferences are based on deep knowledge and experience.
2.2.4 You may want to “mostly follow” because you generally like / admire the other person, or (conversely) to “mostly lead” due you dislike the other person
Let’s start with an example of what I call “liking / admiring”, closely related to “prestige” (see Section 2.3 below), and why someone might want to “mostly follow” in such a circumstance.
Example scenario: {Alice = pop icon, Beth = enthusiastic fan-girl}: I’m Beth, a teenage fan-girl of famous pop singer Alice, whom I am finally meeting in person. Let’s further assume that my demeanor right now is “confident enthusiasm”: I am not particularly worried or afraid about the possibility that I will offend Alice, nor am I sucking up to Alice in expectation of favorable treatment (in fact, I’m never going to see her again after today). Rather, I just really like Alice! I am hanging on Alice’s every word like it was straight from the mouth of God. My side of the conversation includes things like “Oh wow!”, “Huh, yeah, I never thought about it that way!”, and “What a great idea!”. And (let us suppose) I’m saying all those things sincerely, not to impress or suck up to Alice.
That’s the setup. Then the question is: why? Why am I acting that way?
My vague answer is: I really like / admire Alice, so I tend to also really like anything that’s associated with Alice, which includes the things that Alice does and says (and wears, etc.). If Alice says that we should go to a certain bar, then that’s exactly where I want to go, because from my perspective, if Alice wants to go to a certain bar, then that must be a friggin’ awesome bar!
It’s sort of a magic trick wherein, as soon as I catch wind of Alice’s object-level preferences, my brain changes its own object-level preferences, to more closely match Alice’s. See §2.6.1 below for a bit more on what might be happening in my admiring brain, or better yet §4.5–§4.6 of my Valence series.
Conversely, here’s an opposite example:
Example scenario: Alice and Beth are 13-year-old classmates assigned to work together on a group project, and Alice thinks Beth sucks. Let’s say Beth is the least popular kid in school, and Alice is moderately popular, and Alice is just dripping with contempt for Beth. From Alice’s perspective, Beth is lame, stupid, and gross, and everything Beth says or does is lame, stupid and gross as well. However, Alice & Beth have been compelled to collaborate on a presentation about ancient Assyria. They are sitting next to each other. Beth suggests making the background red, and Alice immediately shoots back “ugh, that’s so ugly, no way”. They make it green. Beth suggests putting the title in the center, and Alice rolls her eyes and says “ugh, that’s stupid, it should be top-left”. And so on.
Thus, Alice is evidently motivated to mostly lead in this interaction.
Alice’s attitude here is (let us suppose) not motivated by the goal of annoying Beth. Rather, it’s the “magic trick” above in reverse: as soon as Alice catches wind of Beth’s object-level preferences, Alice’s brain changes its own object-level preferences, in the direction of disliking whatever Beth likes. After all, if Beth likes something, then it must be super lame, from Alice’s perspective. And this applies even to stupid things that it makes no sense to dislike: imagine Alice saying to herself “Ugh, I can’t believe I have to breathe the same air as Beth”, or “Ugh, I hate the stupid way that Beth walks” (and then Alice starts self-consciously walking in a different way from Beth) (related: “bitch eating crackers”). As before, if you’re interested in the mechanism of what’s happening in Alice’s brain when she starts walking funny, it’s just the mirror-image of the above, and hence see §4.5–§4.6 of my Valence series.
(Both the examples in this section are invoking 13-year-olds, because I think kids tend to display social behaviors in a really exaggerated and ridiculous way that makes it easy to notice. I think these phenomena absolutely exist in adults too, but we’re more likely to be hiding and rationalizing it, rather than proudly flaunting it.)
2.2.5 You may want to “mostly follow” to avoid guilt, blame, and/or responsibility if things go badly, or (conversely) to “mostly lead” to get credit if things go well
Let’s start with why someone might want to “mostly follow”.
Example scenario: {Alice = confident person, Beth = nervous person}: I’m Beth, an employee at the widget factory. My fellow-employee Alice and I have to carry a heavy, delicate, expensive machine through a narrow doorway. I’m extremely nervous about this, whereas Alice seems confident. If Alice starts “mostly leading” and making decisions about how to lift and maneuver the machine, then I will breathe a sigh of relief, perhaps related to my subconscious expectation that if something goes wrong, most of the external blame and internal guilt will fall upon Alice, not upon me. Relatedly, in this situation, I will “follow” in various subtle ways, like saying “I’m not sure”, using body language to encourage Alice to take the lead, agreeing with Alice’s tentative proposals, and so on.
I don’t think this example is the same as any of the previous subsections —I don’t need to respect or admire Alice for this to happen, and indeed, I think this vignette is still plausible even if I think that Alice is objectively no more knowledgeable about moving the machine than I am. Actually, to be honest, if it were me personally in this situation, I would probably be “mostly following” even if I thought Alice was somewhat less knowledgeable than me about how to move the machine![1] (Of course, if I thought Alice was dramatically less knowledgeable than me, then I would summon up the courage to lead more—or at least, I sure hope I would!)
How about the opposite?
Example scenario: {Alice and Beth both have good ideas}: The CEO is in town, and emails Alice and Beth asking for a restaurant suggestion. Alice and Beth each have a favorite pizza restaurant. Moreover, Alice and Beth have both frequently suggested their respective favorite pizza restaurant to friends and family, and always gotten rave reviews back. So in this situation, Alice and Beth might both find themselves motivated to be the one who replies to the CEO with their restaurant suggestion, knowing that the CEO will have a great meal and then feel grateful towards the person who suggested that restaurant. Objectively, maybe the CEO shouldn’t feel especially grateful towards the person who made the restaurant suggestion, because (counterfactually) the other person would have made an equally-good suggestion. But that’s not how people think. (Related post.)
I think this is a little slice of what it looks like to be (probably-unconsciously) motivated to “lead” by the prospect of getting credit.
2.2.6 You may want to accommodate (or frustrate) the other person’s preference to mostly lead or mostly follow
Since leading is zero-sum (see §1.3 of the previous post), there’s a push-and-pull thing where if Alice is trying to lead more, then Alice is necessarily also trying to make Beth lead less, and vice-versa. So if Alice expects Beth to have preferences and likely reactions to Alice’s pushiness, those expectations will feed into Alice’s decisions and behavior.
The classic dynamic in this category is “dominance” (more on which in §2.3 just below), where Alice wants to mostly lead for any of the other five reasons in §2.2.1-§2.2.5, and then Beth mostly follows out of fear of angering Alice. But there are many other variations too. For example, maybe Beth mostly follows Alice out of a feeling of compassion rather than fear. Or maybe Beth just wants to frustrate Alice’s desire to lead, for its own sake, out of spite.
2.3 Dominance-flavored versus prestige-flavored interactions
I briefly mentioned the dominance-vs-prestige dichotomy in the previous post (§1.2.1); for a friendly intro, see Dual Strategies Theory (wikipedia), or Elephant in the Brain, or the following very brief summary:
If you hate your boss, but you do what she says anyway because she’ll fire you if you don’t, that’s dominance. If you’re very respectful to a police officer because he has a gun and you don’t, that’s dominance too. Principals have dominance, parents have dominance, psychiatrists keeping you in a hospital against your will have dominance. Prestige is different. A rock star has prestige. He can’t hurt you. You don’t necessarily need anything from him. But you still want his autograph, want to meet him, maybe want to sleep with him. Star athletes have prestige. Actors and actresses. Good bosses who you work hard for not because you’re afraid of them but because you don’t want to let them down. Your parents, if you do what they say out of respect/love and not out of fear of punishment. Heroic leaders like George Washington (except more alive). —Scott Alexander
I propose to relate dominance-vs-prestige to the previous section as follows:
In a typical dominance-flavored interaction:
Alice and Beth both want to “mostly lead”, for any of the reasons in §2.2.1–§2.2.5 above;
Alice does in fact “mostly lead”, but Beth instead “mostly follows”, because Beth is afraid of offending Alice.
In a typical prestige-flavored interaction:
Alice might or might not want to “mostly lead”;
Beth wants to “mostly follow” (or equivalently, Beth directly wants Alice to “mostly lead”), because Beth likes / admires Alice, per the discussion in §2.2.4 above.
As should be clear from the previous section, I am thinking of “dominance” and “prestige” as two caricatured scenarios, rather than a clean sharp division of the universe of status behaviors. For example, if Beth is sucking up to her boss Alice in a calculated, manipulative bid to curry favor, I would say that Beth is “mostly following” for a reason that matches neither dominance nor prestige. Ditto if Beth is “mostly following” in an attempt to avoid blame (§2.2.5 above).
That said, the “dominance” and “prestige” interaction archetypes are important, and I will elaborate on them below in §2.5 and §2.6 respectively.
2.4 “Status” as a tacit shared understanding of how much someone can lead without getting pushback
Let’s circle back to “Making yourself small” by Helen (2018), the blog post featured in §1.2 of the previous post. First, Helen explains what “making yourself big / small” means—as discussed earlier, her “making yourself big” translates to my “mostly leading” and her “making yourself small” translates to my “mostly following”. Second, she defines “status” as follows:
High/low status is about (among other things):
How much power you have
How much attention you can expect
How much space you are entitled to
Making yourself big/small is about (among other things):
How much power you are exercising
How much attention you are demanding
How much space you are taking up
She gives helpful examples of all four quadrants—see the images and surrounding text in the article. Or if you’re in a hurry, here are four examples:
High status person trying to mostly lead — a confident leader.
High status person trying to mostly follow — a really good therapist or tutor, exuding confidence and grace, but also giving you lots of space, encouragement, and deference.
Low status person trying to mostly follow — an unpopular nerd cowering in the face of a bully.
Low status person trying to mostly lead — that same nerd, obviously terrified and trembling, nevertheless stands up and faces the bully, declaring that he’s not going to take it anymore.
Helen also points out in a helpful comment that, in popular discourse, when people talk about body-language and other signs of status, they typically don’t distinguish between actual “status” versus “mostly leading”. (No surprise, since they tend to go together.) For example, in this list, I think “giving or withholding permission” is about leading, whereas “moving comfortably and gracefully” is mostly about “status”.
2.5 More on dominance
2.5.1 You can “lead too much” even without “leading” on an absolute scale
Recall the weighted-average toy model from §1.3 of the previous post, where there’s a range from “0% leading” to “100% leading”, and the two parties to an interaction need to add up to 100%, by definition. Thus,
If Alice and Beth both want to be 60% leader, that’s a problem, because . Maybe that looks like a stereotypical play for dominance.
If Alice wants to be 90% leader and Beth wants to be 20% leader, then that’s also a problem, because . In this case, Alice may well feel like Beth is being presumptuous, despite the fact that Beth is nowhere close to bidding to overtake Alice.
For the second bullet point, you might be imagining Alice as a terrible tyrannical dictator. But you shouldn’t assume that. To illustrate, here’s a sympathetic-to-Alice example:
Let’s say Beth is Alice’s mother, and Beth gingerly mentions to Alice that she’d like Alice to please break up with her girlfriend. And then Alice immediately snaps back “What the heck, Mom! I’m a 30-year-old adult, and I’ll go out with whoever I damn well please!!” Beth is bidding for maybe 10% leadership in this interaction. (If it was much higher than 10%, then Beth would be firmly making a demand, rather than gingerly making a request.) But Alice feels entitled to exactly 100% leadership in making decisions about her own personal love life. We see the resulting conflict—and for what it’s worth I would totally take Alice’s side.
Anyway, I bring up the two bullet points above because people sometimes invoke a kind of “ladder”, or the chicken-derived “pecking order”, model for status, in which the first bullet point is frequently discussed (whose status is “higher” or “lower” than whose?), while the second bullet point is downplayed or forgotten. But my current guess is that “50% leader” is not a special magical threshold that holds any particular importance. At least, not for humans—I don’t know much about chickens.[2]
2.5.2 Alice “takes offense” when Beth is tries to lead more in interacting with Alice (and thus forces Alice to lead less) than Alice wants and expects
Similar to the discussion in The Nature of Offense (Wei Dai, 2009), if someone wants and expects to lead more than the other person is allowing, the first person may describe themselves as “offended”, and describe the other person as “rude”.
(As in the previous subsection, this is relative, not absolute—you can be offended because you wanted and expected 80% leadership but the other person is only offering 60%, or you can be offended because you wanted and expected 30% leadership but the other person is only offering 10%, etc.)
For example, Alice might feel offended if Beth says something inconsiderate, or if Beth is excluding Alice from a conversation that Alice wants an opportunity to be part of, or if Beth is ignoring how Alice might react to a situation. All of those are cases where Beth is not trying very hard to anticipate and accommodate Alice’s opinions and preferences—or in other words, they are all cases where Beth is “leading too much” in the Beth-Alice relationship, from Alice’s perspective (see §1.2 of the previous post).
2.5.3 Being “passive aggressive”: how and why
Let’s say Beth repeatedly borrows Alice’s clothes without asking. And Alice takes offense—i.e., Alice thinks that Beth is “leading” more than she should be in this interaction.
The natural response by Alice is to shift upwards on the “pushiness” scale of the previous post (§1.4).
If Alice dials up her pushiness a lot, we get direct aggressiveness: “Sorry Beth, they’re my clothes, I won’t let you take them.”
But if Alice dials up her pushiness just a little bit, then we get passive aggressiveness: “Hey Beth, I really needed my red blouse last night, but I couldn’t find it, do you by any chance know where it was?”
Why would someone want to be passive-aggressive instead of direct-aggressive? Because in most cases, you want to be pushy enough to get what you want out of the interaction, and no more. (There are some exceptions, including “I feel angry and am lashing out” and “I (consciously or unconsciously) want to establish a reputation as a person not to be trifled with”.) For example, acting more-pushy-than-necessary can give you a stronger reputation as a jerk, which you probably don’t want. As another example, if Alice is a manager, she might use passive-aggression when pressuring her underlings to work late, as a (conscious or unconscious) strategy to avoid full common-knowledge responsibility / blame for that outcome (see §2.2.5 above), especially if upper management is in favor of work-life balance.
Conversely, just as Alice benefits from being passive-aggressive rather than direct-aggressive towards Beth, by the same logic, insofar as Beth’s preferences are opposite Alice’s, Beth may (paradoxically) wish that Alice would be more directly aggressive towards Beth. For example, Beth might be implicitly thinking “if only Alice would only be more directly aggressive, then it would be really easy for me to convince third-parties that Alice is actually the jerk that I know her to be”. Or for the manager vignette above, the underling might think “if Alice would just come out and explicitly demand that we work late, in so many words, instead of insinuating it with plausible deniability, then I could easily strike back by mentioning that fact to the CEO, because the CEO supports work-life balance.”
2.5.4 How do dominance-flavored “status” expectations get set and adjusted?
In §2.4 above, I talked about “status” as a tacit shared understanding of how much a person can lead in a certain interaction context, without getting pushback. Where does this “status” come from, in the case that it mostly involves dominance?
We should notice that this is a game theory problem, almost straight out of Thomas Schelling Strategy of Conflict:
We have a “cooperative game” (each player benefits from having the same “tacit understanding” as the other player, so as to avoid mutually-damaging fights and animosity);
…and the game has multiple possible Nash equilibria (the tacit shared understanding could say that Alice is entitled to X% lead, for any possible X from 0% to 100%);
…and the players are very limited in their ability to honestly communicate (because they would prefer different equilibria, so are incentivized to bluff; relatedly, they don’t necessarily know if a threat is credible without mutually-damaging conflict).
This game plays out on both evolutionary and within-lifetime timescales. On the evolutionary timescale, I think many animals have evolved instincts to “size each other up”—i.e., estimate the ability of a rival to win in a fight, based on its size, strength, etc. I think humans have such instincts too, and that this is a big part of why tall people tend to earn more, win at politics, and so on. I’m not sure exactly how that works in the brain.
Evolutionary game theory is also probably also involved in territorial instincts in animals (i.e., marking a territory and “leading more” / being more aggressive when inside it), but I don’t know the details and am not sure if it’s applicable to humans. (Maybe it is, in that we’re a bit more comfortable / confident in a familiar environment.)
Next we turn to the within-lifetime-learning aspects of this game:
If I’ve interacted with someone lots of times in the past, then we’ve probably settled into an equilibrium (shared tacit understanding of relative “status”), based on our past history of interaction (including antagonism and threats), and deep knowledge of each other and how we’re likely to act. Incidentally, once that equilibrium is established, it may be difficult to change without starting a big fight.
By contrast, if I’m getting to know someone new, then that’s a trickier game to navigate, especially if we’re both trying to avoid animosity. I think one of the factors at play is that we’re both subconsciously pattern-matching our relationship to our previous relationships, along with cultural norms and so on. For example, if I were sitting on an airplane next to Kim Kardashian, I would expect her to feel entitled to “mostly lead” without pushback, just because that’s my sense of the normal dynamic for interactions between celebrities and non-celebrities in public. Therefore, if I want to avoid drama (which I always do), I would probably be unusually deferential towards Ms. Kardashian. This is true despite the fact that I don’t give a crap about her, feel no desire to impress her, have nothing to gain from her, etc.
(See also: “Schelling point”, “The Intelligent Social Web”).
I think the so-called “status-regulation-slapdown emotion” (see here or here) might have some relation to this social / cultural mechanism—maybe people are implicitly thinking something like: “if Beth is mostly leading in some other context, then maybe Beth will get in the habit of mostly leading more generally, including when interacting with me, and I don’t like that”. (The “status-regulation-slapdown emotion” is probably also related to a different dynamic that I’ll get to in §2.6.4-§2.6.5 below.)
2.6 More on prestige
2.6.1 What does “Beth likes / admires Alice” mean, and why does that impact Beth’s behavior in this particular way?
See §4.2 of my valence series for a more extensive discussion and set of examples, but here’s a summary.
As discussed in §2.2.4 above, if Alice is interacting with Beth, and Beth “likes / admires” Alice, then it fills Beth with a motivation to have Alice lead more (on the margin). We might describe this as Alice having prestige-flavored “status”, i.e. an ability to “mostly lead” without pushback from Beth, at least in this specific interaction context.
I’m using the term “likes / admires” here, but it’s not a perfect term. For example, if “Bob likes Alice” as a potential sex object, then that doesn’t count—it’s not what I’m talking about. And if King Henry grows to like and trust his Loyal Advisor Jeeves, then we would probably not say that “King Henry admires Jeeves”, but it is an example of what I’m going for. The terms “Beth thinks highly of Alice” or “Beth respects Alice” also have some overlap with what I’m going for, but also miss the mark in other cases.
In §2.2.4 above, I talked about a “magic trick” where, as soon as Beth catches wind of Alice’s object-level preferences, Beth’s brain changes its own object-level preferences, to more closely match Alice’s. (See §4.5–§4.6 of my Valence series for a bit more about how this magic trick works in the brain.) So an essential precondition is that Beth is thinking about Alice’s desires—and not just abstractly, but as a focus of attention. Perhaps we should say “Beth likes Alice in Alice’s capacity as a person with beliefs and desires and agency”. And now it should be clear why “Bob likes Alice as a potential sex object” is the wrong kind of “liking” for prestige. Ditto for “I like the Phillie Phanatic”.
(Neuroscience tangent: when I say “Beth is thinking about Alice’s desires—and not just abstractly, but as a focus of attention”, what I’m really thinking of is “transient empathetic simulations”, which I believe play a starring role in almost all human social instincts.)
Incidentally, parallel to §2.5.1 above, liking / admiring is a matter of degree, not “positions on a ladder”. If King Henry has trust / respect for his Loyal Advisor Jeeves, then maybe Henry wants Jeeves to be 5% leader, rather than the normal 0.1% expected for his lowly servants. Henry has some motivation to know about Jeeves’s preferences and be pulled by them, but still not much motivation on an absolute scale.
2.6.2 What controls how much Beth “likes / admires” Alice?
“Beth likes / admires Alice” is a special case of Beth’s brain assigning “valence” to concepts in her world-model—a ubiquitous and central part of mental life. I have some generic discussion in §2.5 of my Valence series. As an example, it helps for Alice to do things that Beth likes for other reasons—e.g. Beth can “like / admire” a pop star who sings songs that Beth finds moving, and if Beth has come to believe that skateboarding is really cool, then Alice can get some liking / admiration by being a good skateboarder.
Also, as an important special case, I think if Alice is directly helpful to Beth, e.g. by doing Beth a favor, then Beth will like / admire Alice a bit more on the margin.
2.6.3 Prestige as a “currency” of reciprocal altruism
Background: in evolutionary psychology, “reciprocal altruism” (wikipedia) is the idea of doing favors to unrelated people in the expectation that they may eventually do favors to you in return.
Immediately above I claimed that, when Alice does a favor to Beth, then Beth will probably like / admire Alice a bit more on the margin. And in §2.2.4 above I argued that, the more Beth likes / admires Alice, the more Beth will feel intrinsically motivated to do things that Alice wants Beth to do, including favors.
So this is a complete setup for reciprocal altruism!
I think there’s much more to reciprocal altruism than just that, but it does seem to be a huge piece of the puzzle. Interestingly, this huge piece of the puzzle seems to pop up as a byproduct of other more basic evolved mechanisms, rather than being specific to reciprocal altruism, although I’m not 100% sure about that.
2.6.4 Counterintuitive fights to give rather than receive
The book Elephant in the Brain talks about Arabian babblers, a bird species who apparently fight aggressively for the “privilege” of doing favors for each other, such as putting themselves in danger to guard the nest, or even shoving food down each other’s throats. Similarly, in human affairs, a group of friends at a restaurant will sometimes fight over who pays. Everyone’s object-level preference is to not pay the bill, but their all-things-considered preference is that they positively want to pay.
What’s going on here? Given the discussion above, it’s straightforward: everybody wants other people to like / admire them, and everybody knows that satisfying someone else’s object-level preferences is a way to make that happen more on the margin.
2.6.5 Self-deprecation and downplaying one’s favors
Continuing with the example above, at the object-level, I don’t want to pay the restaurant bill. But if I care less about money and more about accumulating prestige, then I do want to pay the bill. But that also means that, if you pay the restaurant bill instead of me, then you are standing in the way of my desires—not my object-level desires, of course, but my social, prestige-hoarding desires. Therefore, if you really want to be nice to me, then you should: (1) pay the restaurant bill, so that you’re satisfying my object-level desire to save money, while simultaneously (2) spinning that payment as a favor that I am doing to you (“Oh, I would be so honored to buy you this meal”), so that you’re satisfying my social desire to do you a favor.
(But wait: if that whole performance—where you pay the bill and then spin it as me doing you a favor—is what “being maximally nice to me” looks like … then you are in fact optimally accumulating prestige by doing exactly that! Right?? The main thing spinning here is my head.)
I’ve also heard that, in hunter-gatherer cultures, someone who brings meat to the group after a successful hunt will aggressively downplay their hunting skill and the quality of the meat—and everyone else will downplay it as well. (Example.)
(Another factor at play might be countersignaling—i.e., “I’m so generous / good at hunting / whatever that I don’t need to flaunt it”—but I think that’s not as important as the above dynamic.)
2.7 Conclusion
A few particularly important takeaways from this and the previous post are:
Many everyday interactions—requesting favors, delegating responsibilities, making joint decisions, conversational turn-taking, and so on—are at least partly negotiations over conflicting object-level preferences.
People’s preferences often get different weights, a phenomenon I call “mostly leading” versus “mostly following” and discussed in the previous post.
There is a wide variety of motivations and dynamics lying behind the fact that someone might be “leading” or “following” in a specific interaction context. Two of those are “dominance” and “prestige”, but we should probably think of those as two caricatures of common interaction types, rather than two “systems” or “ladders” or whatever. Relatedly, many important motivations like jockeying-to-avoid-blame and manipulative-sucking-up are connected to leading / following but not particularly related to either dominance or prestige.
There are probably evolved innate drives involved in these phenomena. I speculated about what they might be in §2.2.2, but am not too sure about the details. I’m expecting something more like an “innate drive to be liked / admired by other people” [I have further discussion of this putative drive in a different post], and an “innate drive to be feared by other people”, which are upstream of prestige-seeking and dominance-seeking respectively. I expect that there are other relevant innate drives too. I hope to write more in the future about exactly what these drives might be, and how they might work mechanistically in the brain.
(Thanks Rafael Harth, Seth Herd, Linda Linsefors, Justis Mills, and Charlie Steiner for critical comments on earlier drafts. Thanks Rafael Harth for his criticism of my first attempt to talk about social status, which I found sufficiently compelling that I basically wound up restarting from scratch.)
- ^
This example seems really intuitive to me, but I’m concerned that some readers have very different inner lives than mine, so I should spell this out in more detail. My psychological makeup seems to involve an unusually strong drive to avoid feeling guilt / regret later on. And I feel much more guilt upon making a mistake myself, than upon deferring to someone else who then makes a mistake. It’s like, if I can tell myself a story after the fact in which there’s someone else to blame, then it takes some of the sting off the guilt, or something like that. I know it’s not rational, but it’s how my brain seems to work.
- ^
As a different point against the ladder model in chickens, I found an amusing anecdote on reddit of someone who claimed that it’s common to find three hens with a rock-paper-scissors “pecking order” loop.
- Social status part 1/2: negotiations over object-level preferences by 5 Mar 2024 16:29 UTC; 115 points) (
- [Intuitive self-models] 4. Trance by 8 Oct 2024 13:30 UTC; 75 points) (
- [Intro to brain-like-AGI safety] 13. Symbol grounding & human social instincts by 27 Apr 2022 13:30 UTC; 69 points) (
- Neuroscience of human social instincts: a sketch by 22 Nov 2024 16:16 UTC; 55 points) (
- [Valence series] 4. Valence & Liking / Admiring by 10 Jun 2024 14:19 UTC; 46 points) (
- 12 Mar 2024 22:27 UTC; 15 points) 's comment on Some (problematic) aesthetics of what constitutes good work in academia by (
- 11 Mar 2024 12:38 UTC; 4 points) 's comment on Evolution did a surprising good job at aligning humans...to social status by (
- 17 Mar 2024 4:54 UTC; 1 point) 's comment on Social status part 1/2: negotiations over object-level preferences by (
I liked Part 2 as well, but wish it covered more of the “status universe”. Ideas for future posts (things that I wish I understood better):
how wealth/income relate to status
old money vs new money
how does status signaling / status symbols work
why do people choose certain things as status symbols
why are some people willing to pay such a heavy cost to obtain status symbols? how/why does this differ by culture?
how does civilization channel people’s desire for social status into “productive” ends (like increasing a nation’s economic capacity and military power, or making intellectual progress)?
can we design better institutions for doing this through explicit understanding of status?
given that people seem to care a lot about status symbols / positional goods, what implications does that have on AI alignment?
morality as a status game and how that affects alignment/safety
philosophy and intellectual progress in general (including AI alignment/safety) as a status game
if AI alignment were to succeed, how does that affect humans’ status games / social interactions? what will human interactions look like afterwards?
Hmm, I think offense is more about status (“a tacit shared understanding of how much someone can lead without getting pushback”) than how much someone is leading or following in a particular interaction. People tend to take offense when their status is threatened, even if they’re not directly interacting with the person offending them, for example taking offense when you hear someone talk trash about you (or even your country, culture, etc.) behind your back. And we can view someone not following you as much as you want/expect as a special case of this, because the other person is implying through their actions that they don’t think your status is as high as you think.
There’s actually a meta-status problem with any group discussion of status, namely that if the group members judge it to have a below average chance of winning a status competition, in whatever sphere of activity they are engaged in, then its members have incentives to block or ignore the discussion.
Or even downplay the group itself, its quality, etc…, if they can’t prevent the discussion, much like hunting groups for meat. This especially applies for group members who perceive themselves to be in the most marginal, low status, cohort.
The core reason is nobody wants to be known as a 100% guaranteed loser, so anyone who already has below average prospects is going to feel extremely sensitive about even the slightest chance of the group losing future status competitions and thus dragging them down even further.
Although this doesn’t apply to the most valuable group members, who presumably view themselves as having above average status, the opposite problem occurs, namely that actually winning a status competition might attract people who are above them into joining, and thus diluting their own influence, or even worse, relegating them to the second tier. (this doesn’t apply if the group is already at the very highest level)
So paradoxically only the ‘middle-class’ members reliably do anything more than empty talking, at least for status constrained issues. Literally everyone else has incentives to talk a big game but also prevent anything decisive.