Strategic interaction in fiction

S1E1, Succession. Tom and Shiv, fiance and fiancee, are chatting about Shiv's dad's (Logan) upcoming birthday.

Tom: Shiv, it's a fucking disaster. I need to strategize my gift. What can I get him he'll love?

Shiv: I don't know, my dad doesn't really like things.

Tom: It needs to say I respect you, but I'm not awed by you. And that I like you, but I need you to like me before I can love you.

Shiv: Look everything you get him will mean equal amount of nothing, so make sure it looks like ten to fifteen grand's worth and you're good.

Jane Austen, as we'll see, believes that couples are the ultimate strategic units: at their best, marked by deep trust, shared exposure/experience, and united in joint interest. Tom and Shiv's problem, as we'll also come to see, is that they're not exactly united in interest or trust. But Tom, at least at this stage, believes they are, hence the strategizing.

The other thing I like about this excerpt is that Tom explicitly frames the gift as a communication. "It needs to say" X and Y. It has a purpose, and that purpose is manipulating their relationship (Tom and Logan's). Tom's longterm goal, as future seasons reveal, is to take over the company as CEO, a position Logan currently holds, and thus their relationship—specfically, Logan's impression and opinion of Tom—is crucial.

Here's one of the great scenes from House of Games, whose motto could be "Everybody gets something out of every transaction."

Here's our setup: Margaret is a best-selling psychiatrist so overwhelmed by work, so frustrated by her helplessness to meaningfully improve her patients’ lives, that she enters a manic state, shows up in a pool hall across town and starts threatening a loanshark who is pressuring a patient of hers. It’s no coincidence that a literal financial debt is the MacGuffin that makes this plot run.

The loanshark, Mike, turns out to be more petty conman than gangster; Margaret’s intrigued. His bluffing and persuasion tactics sound like a great idea for her next bestseller. She asks him to show her his moves. It’s clear what she’s getting out of it—another book, excitement in her life, a model of masculine assertiveness. (Mike: “I think what draws you to me is this: I’m not afraid to examine the rules and to assert myself.”) But it’s not clear what he’s getting out of it, and Margaret never stops to ask. This is her first major error.

Mike takes her to the Western Union pictured, pulls a con on a wholesome, off-duty marine who needs a bus ticket to Camp Pendleton. The con works by presenting false information—after all, values are hard to alter, but it's priors that regulate action. At the last minute, when the marine’s practically begging Mike to take his money, Mike walks away, tells him to keep it. Remember Schelling: small series of trust-building exchanges are “practice” for the real thing. Why would Mike skip out on the money; he’s a criminal. Because he doesn’t want to appear that way, as a bad guy. He’s gaining Margaret’s confidence. The marine’s not the real mark. She is.

Eventually and predictably, the bestselling psychiatrist (hat-tip to Game A) gets scammed out of $80k in cash, believing, under false pretenses, that it’s a mob payoff—that the payment, in cash, handed over to Mike, will save her life. How could she be so gullible? He told her what she wanted, then he gave it to her. She was getting everything she ever wanted. The real question is: Why would she ask questions?

This is exploitative manipulation, in the "all communication is manipulation, but some is mutually advantageous and some is exploitative" sense. It isn’t exploitative because he lied to her, or because his intentions were “bad.” It’s exploitative because their relationship took more from her than it delivered in turn, and because it was designed, deceptively, to do so. The final scene, where (spoilers) Margaret shoots Mike to death in an abandoned airport hanger, marks the end of their ongoing transaction—the part where they “settle up.” The final scenes show her so healthy and self-contented she’s practically glowing. She’s taken life into her own hands—asserted herself. It’s a comedy because it ends happy, and it ends happy because she walks away from it better than she started.

Venkat's "Gervais Principle" speculates that an org tends to have three kinds of employees: sociopaths, the clueless, and losers. Sociopaths are in a constant battle for power and dominance. The clueless are "company men" who genuinely believe in the organization and are too stupid to think critically about their situation. Losers are those at the bottom who make a straightforward exchange of time for money while still looking out for their self-interest, they tend to find meaning outside the organization, and the most enlightened among them do the minimum work possible to keep coasting.

These three groups talk to each other in different dialects: powertalk, posturetalk, babytalk, and straight-talk. Powertalk occurs whenever sociopaths encounter each other. Sociopaths see power in every interaction, and thus every conversation and exchange of information between them leads to power shifts, and must be controlled meticulously to maintain maximum power.

What distinguishes Powertalk is that with every word uttered, the power equation between the two speakers shifts just a little. Sometimes both gain slightly, at the expense of some poor schmuck. Sometimes one yields ground to the other. Powertalk in other words, is a consequential language.

Babytalk is equally manipulative, but is done asymmetrically, to those clueless who are unable to speak powertalk or detect when babytalk is being spoken to them. Enlightened losers and high-up sociopaths both speak babytalk to clueless managers.

Posturetalk is a clueless attempt at powertalk. Straight talk—what @amirism might call a straightforward discussion of the "rules of the game"—can only occur between individuals who have nothing to lose or gain: "It is the ordinary (if rare) utilitarian language of the sane, with no ulterior motives flying around. The mean-what-you-say-and-say-what-you-mean stuff between two people in a fixed, asymmetric power relationship, who don’t want or need to play real or fake power games."

Venkat goes on to give an example from The Office of fluent powertalk, here between enlightened loser Jim and sociopathic powerhouse David Wallace.

At a Dunder-Mifflin management party, shortly after Michael and Jan disclose their affair to David Wallace, per HR requirements, Wallace casually invites Jim to blow off the party for a while and shoot hoops in the backyard. Once outside, Wallace nonchalantly asks, “So what’s up with Jan and Michael?” He is clearly fishing for information, having observed the bizarre couple dynamics at the party.

Jim replies, “I wouldn’t know…(pregnant pause)…where to begin.” (slight laugh)

David Wallace laughs in return. This is as eloquent as such a short fragment of Powertalk can get. Here are just some of the messages being communicated by the six words and the meaningful pause and laugh.

1. Message 1: It is a complex situation (literal).
2. Message 2: I understand you think something bizarre is going on. I am confirming your suspicion. It is a bizarre mess, and you should be concerned.
3. Message 3: This is the first significant conversation between us, and I am signaling to you that I am fluent in Powertalk.
4. Message 4: I know how to communicate useful information while maintaining plausible deniability.
5. Message 5: I am not so gratified at this sign of attention from you that I am going to say foolish things that could backfire on me.
6. Message 6: I am aware of my situational leverage and the fact that you need me. I am not so overawed that I am giving it all up for free.
7. Message 7: I am being non-committal enough that you can pull back or steer this conversation to safer matters if you like. I know how to give others wiggle room, safe outs and exits.
8. Message 8: You still have to earn my trust. But let’s keep talking. What do you have that I could use?

Crucially in powertalk, you play with real table stakes; "the currency is most often reality-information." (i.e. epistemic capital, as valuable as symbolic capital) The other dialects (Venkat claims) play without stakes. I don't quite understand this, because posturetalk and babytalk both seem to attempt, manipulatively, to bring about realities (i.e. there are stakes, the speakers are trying to make something happen). They're just performed by, or received by, the incompetents/clueless.

(Sidenote: I've been wondering if it'd be useful to drag in Carse's finite vs infinite games, because it doesn't seem like a coincidence that this kind of interaction happens so much in business and high society (that is, over literal and symbolic capital). There are ways of being that seem more cooperative and less full of conflict, where strategic interaction looks very very different—more like improv, less like warfare.)

Theory of mind is one of the biggest skills that makes a sociopath a sociopath—though, as Jane Austen illustrates in her fiction, there is a blindspot insofar as high status people do not actively model the mentalities of low status people, which leads them to misunderstand their subordinates' motivations. So is context-sensitivity: the clueless, in Venkat/Ricky Gervais's formulation, fail in part by following "formulas" for behavior, without an artfulness in deploying them. Thus, in a salary negotiation, Michael Scott "prints off negotiation tactics from Wikipedia and attempts to use a series of recommended formulaic tactics. First he tries switching chairs and rooms to disorient Darryl. He merely disorients himself. Next he tries to follow a rule 'not to be the first to speak.' Sadly, he can't stand the tension, and oblivious to the irony, breaks the silence with 'I will not be the first to speak.' At which point Darryl calmly comes back with, 'Alright, I can start.'" In other words, like all human social assessments (cf flirting), the symbols or metonyms that can encode or signal deeper qualities, the markers of power must either be hard-to-fake, either by being costly or being anti-inductive.

Why can’t you learn Sociopath tactics from a book or Wikipedia? It is not that the tactics themselves are misguided, but that their application by non-Sociopaths is usually useless, for three reasons.The first is that you have to decide what tactics to use and when, based on a real sense of the relative power and alignment of interests with the other party, which the Losers and Clueless typically lack. This real-world information is what makes for tactical surprise. Otherwise your application of even the most subtle textbook tactics can be predicted and easily countered by any Sociopath who has also read the same book. Null information advantage. The second reason is that tactics make sense only in the context of an entire narrative (including mutual assessments of personality, strengths, weaknesses and history) of a given interpersonal relationship. The Clueless have no sense of narrative rationality, and the Losers are too trapped in their own stories to play to other scripts. Both the Clueless and Losers are too self-absorbed to put in much work developing accurate and usable mental models of others. The result is one-size-fits-all-situations tactical choices which are easily anticipated and deflected.

Note: both of these reasons can be summed up by "context-sensitivity." This is a real cognitive skill! It takes serious computation! In that vein, I'm tempted to say that winning in anti-inductive games is not so much a costly signal as it is an indice, much like the depth of an elephant seal's roar is directly, causally, and unfakeably correlated with body size.

Also interesting from "Gervais Principle," on spirit vs letter:

So effective Sociopaths stick with steadfast discipline to the letter of the law, internal and external, because the stupidest way to trip yourself up is in the realm of rules where the Clueless and Losers get to be judges and jury members. What they violate is its spirit, by taking advantage of its ambiguities. Whether this makes them evil or good depends on the situation.

(Some of the ideas here come from Michael Suk-Young Chwe; I'll just add a MSC citation after those lines.)

Austen has a number of names, in her novels, for different parts of strategic interaction: "foresight" (in the sense of prediction and simulation), "sagacity," and "penetration" (in the sense of seeing through appearances, to reality) [MSC]. Sense and Sensibility's Elinor "possessed a strength of understanding, and coolness of judgment"—clear perception, unbiased by emotions, being two of the "gamesman" qualities Goffman singles out in his own Strategic Interaction.

And strategic maneuvering, especially in matters of courtship and finance (tied up as they are in matters of inheritance and dowry), is the foundation of Austen's drama. MSC:

Lucy Steele, a distant cousin of Lady Middleton, confides in Elinor [who has had somewhat of a tryst with Edward] that she has been secretly engaged with Edward for four years. Elinor easily understands Lucy's strategic objective in telling her so: "it required no other consideration of probabilities to make it natural that Lucy should be jealous; and that she was so, her very confidence was a proof. What other reason for the disclosure of the affair could there be, but that Elinor might be informed by it of Lucy's superior claims on Edward, and be taught to avoid him in the future?"

Later, Elinor reveals to her sister Marianne that, on receiving this information, she realized quickly that Lucy's "suspicions" had to be "opposed," principally by "endeavouring to appear indifferent where [she was in actuality] most deeply interested." MSC calls this "self-command" on Elinor's part, a strategic ("sensible") restraint in contrast with Marianne's cult of "sensual" sensibility, her romantic inclination toward emotional expression at all times. Austen, as I read her, takes the side not of authenticity—as romantic worldviews do, with their worship of "sincerity" and self-expression as ends in themselves—but rather sees authenticity or inauthenticity as instruments which can be used productively or unproductively as ends.

One of the principle biases affecting characters in Austen's novels is that of "eagerness of mind" (as S&S's Marianne is attributed). It is an overactive imagination [MSC] and a desire to believe some preferred state of affairs rather than face reality. The two other primary strategic impediments are what MSC calls being a "strategic sophomore"—in short, being overly confident in one's strategic abilities—and being naive or "clueless," in the same sense Venkat intends, of uncritically taking public appearances as genuine reality. MSC:

Mrs. Dashwood's belief in Willoughby is based largely on his manner and appearance: "Has not his behaviour to Marianne and to all of us, for at least the last fortnight, declared that he loved and considered her as his future wife?"

When an overactive imagination and sophomoric overconfidence combine, we see characters like Emma Woodhouse or Marianne Dashwood making inappropriately sized inferential/abductive leaps from their observed information. Thus, when Colonel Brandon in S&S tells a story of a woman of whom Marianne reminds him, Elinor (Austen's ideal strategist) "connect[s] his emotion with the tender recollection of past regard," but does not attempt—since she has no evidence to base it on—any further speculation as to in what way he tenderly regards this woman, or the specifics of their history. It is enough, for her, to have firm reasons for believing only that the regard is tender, and their history meaningful. Marianne, meanwhile, "would not have done so little," Austen writes—"The whole story would have been speedily formed under her active imagination." This is a species of epistemic humility that we might call abductive humility, perhaps somewhat related to Sarah's concept of "indexical geniuses," who refuse to speculate into generalizations. (We might say Elinor is in a Goldilocks zone of abductive ambition, whereas Marianne is too hot and Sarah's indexical geniuses are too cold.)

Who is Greg, in Succession? A cousin of the primary family, strategically clueless, hopelessly naive, and somehow promoted up the ranks of the family company by upper exec Tom. Here's a good portrait of his personality (watch him posturetalk Logan ninety seconds in):

What do we already know about him, without getting into details? From Goffman comes the concept of players, parties, and pawns. Players play on behalf of a party (e.g. an ambassador plays for a country) but may also be their own party in a given situation. Pawns also work on behalf of a party, but their participation is more passive than active: they are used, often at personal cost and harm, to further the party’s goals. Players typically act on a party’s behalf because it is also, simultaneously, in their own interest (up to a point). No similar alignment exists for pawns, although they may be tricked into believing it does.

From Venkatesh Rao we get the phenomenon by which sociopaths, that is, strategic, selfish powerplayers, at high levels of organizations, promote “clueless” (strategically incompetent) players to middle management positions, so that they can be manipulated and, if a project goes south, take the fall on behalf of the sociopath. This is a kind of cronyism, the cronyism of sociopaths to the clueless: the sociopath gains cronies who will support him unconditionally, and uncritically implement or further his agendas, while also gaining a fall-guy.

And we get a similar dynamic in Jane Austen’s Emma. Wealthy, high-class protagonist Emma Woodhouse rears orphan Harriet Smith, hoping to give her some necessary cultural and etiquette instruction and improve her marriageability. The relationship isn’t quite as exploitative as the sociopaths’ toward clueless cronies, but there is a similar logic: Emma gives Harriet greater power and agency, and since they are allies, this increases Emma’s power in turn, since Harriet is now both more indebted to her and more powerful.

Greg ends up taking the hit so hard he has to testify in front of Congress:

"Can't make a Tomelette without breaking some Greggs," Tom jokes.