There is probably no contrast more striking, in the comparison of the mixed-motive and the pure-conflict (zero-sum) game, than the significance of having one’s own strategy found out and appreciated by the opponent. Hardly anything captures the spirit of the zero-sum game quite so much as the importance of “not being found out” and of employing a mode of decision that is proof against deductive anticipation by the other player. Hardly anything epitomizes strategic behavior in the mixed-motive game so much as the advantage of being able to adapt a mode of behavior that the other party will take for granted.

Thomas Schelling

Previously.

If each player’s future outcome (and optimal course of action) is dependent on other players’ courses of action, then each player is in turn strategically incentivized to manipulate his fellow players’ models of himself, in a bid to alter their behavior in ways amenable to his own preferred outcome. This can be perfectly honest—as in notifying a dinner party host about dietary restrictions, to alter the course of meal preparation—or perfectly dishonest—as in misrepresenting one’s income on a first date to increase odds of a sexual encounter. From here on out we can just call this process of strategically self-representing “impression management” (following Erving Goffman’s phrase).

Insofar as the components of strategy—relevant to its processes of reading and writing impressions—are agent choice and preference, capacity and belief, many of the larger strategic patterns employed by players involve modifying, constraining, or selectively revealing such components. One’s preferences or values make up the “theological” dimension of choice—what “ought” to be the case—while one’s capacities and beliefs make up the “logistical” dimension—what is the case, and what is possible or likely given the case. I’ll talk here of strategic legibility vs illegibility, strategic commitment vs non-commitment, and strategic ignorance vs. knowingness—three broad axes by which individuals manipulate others’ models of the players. 

Legibility vs. Illegibility

Legibility, here, refers to a certain transparency, or public availability, of information which it would be theoretically possible for a player to conceal: one’s desires and preferences, one’s emotions and attitudes, one’s patterns of behavior and belief, an interior life generally. It is the quality of being easily “readable” or interpretable. Many of the advantages of legibility are contained already in the discourse around personal “authenticity”: a reputation for honesty, the lower cognitive overhead of keeping a single set of books.

Illegibility, meanwhile, sees synonyms like inscrutable or opaque. An illegible player’s motivations and attitudes are kept private, and observers find his behavior difficult to predict. A player whose beliefs about the world, capacities to act on the world, and desires from the world are all legible, is a player whose next move is similarly easy to predict. Sontag, in “Aesthetics of Silence,” discusses the way that asymmetric self-disclosure reflects and reinforces asymmetric power relations in Ingmar Bergman’s Persona:

For a person to become silent is to become opaque for the other… The ways in which this opaqueness induces anxiety, spiritual vertigo, is the theme of… Persona. The theme is reinforced by the two principal attributions one is invited to make of the actress’ deliberate silence. considered as a decision relating to herself, it is apparently the way she has chosen to give form to the wish for ethical purity; but it is also, as behavior, a means of power, a species of sadism, a virtually inviolable position of strength from which to manipulate and confound her nurse-companion, who is charged with the burden of talking. 

Here, immediately, we are already coming to see that legibility is a more trusting, pro-social approach, while illegibility is seen as defensive or downright hostile (because it is best suited to adversarial, rather than cooperative) games.

Of course, legibility can be feigned—one can appear to be an “open book” while in reality “keeping cards close to the chest.” We can call this strategy pseudo-legibility.

Coordination vs. Conflict

And it is worth stressing: there are consequences to being known. Each piece of additional information held by one’s rivals improves their strategic situation. To possess accurate information—to have true beliefs—is almost always advantageous. To be predictable, in a situation of conflict, is almost always to be disadvantaged. If, for instance, it is well known that as a tennis player, I hit to my opponents’ backhands more often than an average player, my opponent will optimize around this historical fact as a prediction about my future behavior. He will move his body’s location on the court in order to better return hits to his backhand, which will make him slightly less able to return shots to his forehand. Thus, it behooves my opponent to know my strengths, for he can now optimize to guard against them, just as it behooves him to know my weaknesses, in order to exploit them. 

(However, when this accommodation of my historical pattern occurs, deviation from history is subsidized. We can call this an “anything but” subsidy, a negative subsidy insofar as it may also be described as a penalty for my performing an expected, and thus preempted, move. This is what makes tennis, like all strategy games, anti-inductive. If I now repeatedly hit his forehand, he will have to adjust again, and then perhaps I can switch to his backhand, and so on.)

But there are also benefits to being known, precisely because we are not, as social animals, embedded in situations of pure conflict, but rather in “mixed-motive” games, where there is always both a shared interest and a diverging interest. The buyer and seller at a trade stall may haggle ruthlessly, but they are both, ultimately, invested in reaching a deal. Even bitter Cold War enemies like the United States and Soviet Union shared an interest in avoiding mutual destruction, and the end of humanity; this shared interest allowed various nuclear crises to be de-escalated.

To get perhaps the most obvious dynamic out of our way, we can generalize that whether legibility, pseudo-legibility, or illegibility is strategically preferable in a given situation has to do with whether, in reality, one gains an advantage from honestly signaling one’s true traits or next move. In deterrence, demonstrations of strength are most effective when made clearly and unambiguously. Their goal is to show the real power disparity between players, to discourage further conflict by appealing to the shared interest among parties. And given two strategically rational, self-interested players, such displays move all parties to a better Pareto equilibrium: in light of the dominant player’s advantage, it now appears a better strategy for the less-powerful player(s) to keep peace instead of quarrel; at the same time, it is preferable to the dominant player not to incur risk or resource depletion in conflict with rivals, whatever the eventual outcome. Although the dart frog is in competition with its predators, or an alpha chimpanzee with upstart rivals, they will often end up advertising, rather than cloaking, that they are poisonous or strong: the goal is not to slay one’s enemies—self-interested players are never malicious—but to ensure one’s own survival, and optimize one’s own future prospects. Other players’ welfare is relevant only insofar as it threatens or infringes on these prospects. 

Honest signaling as deterrence is one way of ensuring cooperation, but more explicitly, we can see that legibility as a general policy creates possibilities for greater interpersonal cooperation—after all, if no one knows your motivations and preferences, how can any one envision and propose mutually advantageous arrangements? How can they trust you? This is, after all, why intimacy is about knowing deeply and being known deeply. Even sex is—is perhaps archetypally so—an instance of incredible simultaneous vulnerability and opportunity, from the genetic perspective. And thus building intimacy looks like a slow process of mutual disarmament, incrementing to avoid total destruction.

(Consider the social role of alcohol, or inhibition-lowering intoxication more generally. Individuals are able to reveal their internal feelings, or make propositions, which make them more vulnerable to later embarrassment and social sanctions—but which also makes possible certain couplings, commitments, or resolutions that otherwise would not have occurred because the necessary self-legibilization was, by the sober mind, perceived to be too risky. If an individual is inebriated enough to plausibly not remember the disclosures in the morning, then the disclosures are further de-risked: both parties can “pretend it never happened,” allowing their relationship to continue without awkwardness or changed status. Different cultural interpretations of whether alcohol gives false feelings, or reveals true feelings, no doubt further changes the relevant calculus.)  

To give an example of the relationship between, on one hand, coordination and legibility, and conflict and illegibility: In traffic situations, players constantly signal their intentions through brake lights and turn blinkers. Generally speaking, the dominant motivation which underlies these interactions, and drives incentives for self-legibilization, is the desire not to cause or partake in an accident—which at best is financially expensive and at worst fatal. This desire so overrides other considerations that in most interactions, intentions are publicly broadcast, with ample advance warning, so that other players can accommodate a lane shift or slow-down. The desire to coordinate dominates, and self-legibilization with it.

In auto racing, avoiding accidents still remains a priority, but miniscule positional advantages over competitors increasingly jockey for priority with safety concerns. Tellingly, neither NASCAR nor Formula 1 cars are built with turn signals, which would add unnecessary weight without benefit. In drafting pairs, leading cars engage in “mirror driving,” attempting to predict on which side the drafting car will make its pass, and thereby pre-empt the pass by cutting the drafting car off. In Formula 1, teams will attempt to conceal, in advance, their pit stop strategy, which if known would open up “blocking” maneuvers by other teams.[1] Here, illegibility of intent is the dominant strategy, with some exceptions.[2]

Familiar vs. foreign

We will take as a premise, in this paper, some minimal attributes of the predictive processing or free energy theories, that living organisms are broadly uncomfortable with ambiguity and uncertainty, and wish to gain predictive power over their environments, so as to increase their own power to realize preferences within it. These epistemic states of partial knowledge, or dueling interpretations, make optimizing behavior difficult for an organism—when two strategic approaches are mutually incompatible, which to choose?—while the presence of “unknown unknowns” can conceal lurking, catastrophic tail risk. Thus, that which is familiar is comfortable, even as it is “boring,” and the foreign is often uncomfortable or even stressful, even as it is “interesting.” As a species, we prefer to reduce the novel by way of pre-existing taxonomy, rather than “staying with the trouble” it creates.

(I’ve argued that in aesthetic encounters, we have a higher tolerance of novelty and expectation-subverting surprisal, because we have some awareness that we are operating within a low-risk “sandbox” rather than the zone of long-term consequences that characterizes “real life.” And yet still, very few fiction readers are interested in experimental literature; even readers of experimental literature are frequently turned off by incoherence or nonsense—that is, information they cannot “make sense of”; Barthes refers to the ability to “eroticize” ambiguity as a trained muscle.)

Thus, in social interactions or public spaces, individuals “actively constrain their own behaviours so as to make themselves more easily predictable by other agents” (Andy Clark, Surfing Uncertainty). This is typically referred to as “prosocial” behavior: making others comfortable by “doing being ordinary.” Playing a well-understood social role is one means of creating an easy interface for others to interact with. (We all know, for instance, the ritual script for interaction with a waiter.) Individuals can better accommodate one anothers’ preferences, avoid conflict, and seek Pareto-optimal organization. 

As James Carse writes, “[A] title has a specified ceremonial form of address and behavior. Titles such as Captain, Mrs., Lord, Esquire, Professor, Comrade, Father, Under Secretary, signal not only a mode of address with its appropriate deference or respect, but also a content of address (only certain subjects are suitable for discussion with the Admiral or the Holy Mother), and a manner of address (shaking hands, kneeling, prostrating or crossing oneself, saluting, bowing, averting the eyes, or standing in silence).” What you say is who you are is who you’re talking to, the content not constrained but created by the form and addressee.

In situations of genuine novelty, analogy or metaphor is the most powerful technology for familiarization, because it de-troubles “trouble” by showing its likeness to known—and thus untroubling—factors. When players attempting to coordinate reach for the same metaphor, they may keep themselves remarkably in sync despite navigating unfamiliar terrain.

Unpredictable action, meanwhile, is usually perceived as hostile or aggressive—which is part of the Asperger’s stigma (a failure not just to model another mind but to incorporate its model of you into behavior). The concept of a “wildcard”—an unpredictable player in a social scene—can be exciting but also anxiety inducing, depending on the stakes and disposition of the observing player. To a watchful, wary observer, oddity is suspicion-arousing precisely because one does not know what to make of it (and thus it could be anything… hence the tail-risk). McAndrew and Koehnke (2016) write that public illegibility is commonly read as “creepy”:

…what exactly is it that our creepiness detector is warning us about? It cannot just be a clear warning of physical or social harm. A mugger who points a gun in your face and demands money is certainly threatening and terrifying. Yet, most people would probably not use the word “creepy” to describe this situation. It is our belief that creepiness is anxiety aroused by the ambiguity of whether there is something to fear or not and/or by the ambiguity of the precise nature of the threat (e.g., sexual, physical violence, contamination, etc) that might be present. Such uncertainty results in a paralysis as to how one should respond.

Paralysis—in other words, predictive uncertainty. “While they may not be overtly threatening,” the authors continue, “individuals who display unusual patterns of nonverbal behavior, odd emotional responses, or highly distinctive physical characteristics are outside the norm, and by definition unpredictable.” The phenomenology of creepy feelings is not certainty that one’s interlocutor is malicious, but an uneasiness as the possibility.[3] Just as people “become uneasy in environments that are dark and/or offer a lot of hiding places for potential predators,” they are equally uneasy around unpredictable agents. Given that being seen as creepy “undoubtedly creates an impediment to comfortable future social interactions with [the creepy] person,” it is also clearly true that those individuals who come across as creepy take a serious social hit—are excluded from many shared games. Thus, even false presentations of pseudo-legibility may be necessary if one wishes to occupy public spaces or coordinate with others.

In a similar vein, strategizing in novel situations is experienced as “stressful, risky, and forbidding” (Vollmer 2013) for many players involved. Garfinkel’s breaching experiments are one of the better-known illustrations of this effect. Explicit negotiation and bargaining not only risks social offense and damaged relationships (see the common taboo against blurring friendship and finance), it also risks players feeling they’ve gotten a bad deal. It is the indeterminacy of novel arrangements that arouses the feeling that a better deal was possible, incurring regret, a sense of injustice, or bitterness. And this in addition to such negotiations taking a non-trivial amount of time to reach an arrangement. Finding ways to lower the social risk, as well as the cognitive and temporal overhead, of doing business becomes paramount. Thus, Schelling writes, players who routinely interact are on a constant “search for stable, mutually nondestructive, recognizable patterns”—the “creation of traditions.” Once a traditional arrangement is discovered—be it standard legal contract terms (and the boilerplate language that expresses them), or else a clever ethnomethod like “I slice you choose”—it is difficult to challenge or dispute, transitioning the interaction from explicit, “open season” bargaining to more bounded bargaining around ritual allowances. Historical precedent can establish behavioral norms, and breaking such norms can become taboo (i.e. reputationally damaging). This preference for predictable, stable interaction is evidenced across everyday life, and makes a great deal of sense given the predictive processing frame. 

(This gives us a fresh perspective on the so-called fallacious slippery slope arguments of conservatives, in opposing breaks from tradition: there is a non-trivial way in which certain practices are considered “open” or “closed” to revision, and any modification “breaks the seal” on a norm that might otherwise be considered, functionally, a fact of nature. The surplus chaos that dramatic social upheaval historically brings with it, as in the case of a revolution, testifies to this problem. While certain coordination schemes may feel “rigged” or lopsided, instead of fair, the absence of coordination will cause all parties to suffer in an absolute sense. The question becomes whether the temporary cost of such disruptions can be compensated by a successful negotiation of the social contract.)

Again, we can see here that, broadly speaking, in situations where aligned goals and coordinative hopes dominate the interactive calculus, legibility is the dominant strategy. In interactions dominated by conflicting desires, illegibility gains the upper-hand. Thus, if someone is making themselves legible—actually legible, and not pseudo-legible—they are likely attempting to initiate cooperative, alliance-building rituals. And this is true even in situations like nuclear deterrence: demonstrations of strength aimed at securing tacit or explicit peace treaties are most effective when made clearly and unambiguously.[4]

Course vs. fine-grained

From an interaction perspective, clothing and language are surprisingly similar systems—partly functional, partly fashionable, a way of communicating “tribe” through “vibe.” To broadcast a set of dispositions is to attract possibilities for coordination: those who appreciate similar cultural objects and experiences, or who share similar values, can locate one another; because they share schemas for signification, they can communicate more reliably, precisely, and effortlessly. And even those from different tribes are able to more easily (if not always as precisely) interact with one another via glorified stereotype. 

The approximate social role which is indicated by an individual’s cultural composite (which includes their clothing, language, aesthetic preferences, etc) in turn creates a set of expectations in his interlocutor—but these expectations, and the ability to make fine-grained role-expectation distinctions, differs between insiders and outsiders. It is well-known that a new musical genre—infamously, in contemporary culture, metal, country, or reggae—can all sound the same to virgin ears. With regular exposure and experience, the listener learns to distinguish, to identify differences between works instead of merely their similarities. “Familiarity,” William James writes in Principles of Psychology, breeds in us discrimination. “Such vague terms as ‘grass,’ ‘mould,’ and ‘meat,’ do not exist for the botanist or the anatomist. They know too much about grasses moulds, and muscles.” Within hippie or hipster subculture, there are many meaningful distinctions or sub-types to draw; to outsiders, there are only hippies and hipsters. 

This has some important implications. If one is trying to “point”—to metonymically signal some quality, some significant reality or position-taking—one will be understood as crudely or precisely, as ambiguously or unambiguously, as one’s interlocutor’s schema allows. We can call this “pragmatic reader-response,” after the school of literary theory which places literary meaning in the interpretation of the reader. The levers available to pull are the only levers which can be pulled. Metaphysically, of course, there is no truth or reality to whether meaning “really” is the speaker’s intent or the listener’s understanding, but pragmatically, for the purposes of the speaker, who is trying to accomplish things with language, to create an impression in his listener or alter his listener’s actions, it is the listener who counts. If the listener’s schema has been associatively trained to link a behavior to hostility or aggression, then to some extent, in the real, objective unfolding of events across the interaction, the speakers’ (e.g. perfectly benevolent) intent is irrelevant except insofar as it can, through multiple coordinating signs, eventually perhaps contradict and overrule the listener’s misapprehensions.

All this is just to say that legibility or illegibility is not always the result of strategy but of translation failures between schemas.

Ignorance vs. Knowingness

Strategic ignorance is one of the better known epistemic strategy this side of the blogosphere, beginning with Schelling’s work on the advantages of not knowing—or, pragmatically, of credibly not appearing to know (since, opticratically, in strategic situations appearance, and thereby the beliefs of one’s rivals, are all that functionally matter for manipulation purposes).

“There is power both in limiting the responses that are available to you and limiting your knowledge,” Sarah Perry writes in “Cooperative Ignorance”—the former often a result of the latter. If one is ignorant of a development, one is unlikely to respond to it. If one is ignorant of situational problems or their possible solutions, one is unlikely to solve them. Knowledge in many cases is a prerequisite of or equivalent to capacity.

In general, possession of knowledge is strategically advantageous, while actual ignorance is strategically dangerous. Here, however, we are more concerned with epistemic manipulation and its appearances—not what is actual, but what one leads one’s fellow players to believe. If a perfect performance confers certain benefits that a poorly communicated reality does not, then the game’s dynamics are fundamentally opticratic. (Along the lines of the discussion of strategic legibility and illegibility above, to honestly self-representation is often to act in a cooperative spirit, while mis-representation is is often a form of defecting or exploitation.)

When one “plays dumb,” one may be trying to defend a friend (who ought not have spread a piece of gossip…), or hide information that one ought not to know (e.g. Googling a first date in advance), or be attempting to catch an interactant in a lie (because he is not aware his conversant has access to the actual information), but in either case, the goals of the interactants are in conflict: one party desires “the truth” of an individual’s state of knowledge, the other desires to present a picture of the truth that is most in accord with his goals. And this is how strategic interaction, like the selection game, generally plays out: the observer, since he will benefit from accurate information, attempts to determine it, while the observed, who benefits from creating an impression in the observer that furthers his (the impression-creator’s) goals, often has an active opposition to the truth being discovered. Goffman distinguishes between “control moves”—moves intended to manipulate one’s opponent through false appearances—and “naive moves”—moves that are performed without regard for their effect on an observer (indeed, moves that may be performed without knowledge of an observer). To feign naivete, while in actuality making control moves, is one form of weaponizing ignorance.

But there are exceptions, cases where an observer does not want the truth, or where it may ultimately seem most cooperative—best for the relationship from both sides—to feign ignorance. It is hard to say whether these situations are “really” marked predominantly by conflict or coordination. Often, the interrogating party may claim to want access to the unfettered truth, but in reality wishes anything but. Sometimes, the interrogating party’s wish for truth is sincere, but they will be happier and more profitable from a lie. In these cases, we may purposefully ignore, or feign ignorance to, mea culpas and party fouls, moral transgressions and boundary crossings, which would threaten the relationship—see, for instance, the plausible deniability of drunken confessions, or of sexual implicature (“Would you like to see my etchings?”).

Since knowledge is a prerequisite for a proper course of action—hence why keeping the enemy in the dark is broadly advantageous, and acquiring as much (high-quality, relevant) information as possible is broadly advantageous too—then in affairs of law and justice, knowledge can be a prerequisite being held responsible for a decision. If one did not know, how could one have responded, or acted otherwise? (“He doesn’t know better,” the mother excuses her four-year-old biting into his playmates arms.) This applies to legal concepts of responsibilities, and the punishments accorded—hence organizational higher-ups have good incentive to stay ignorant to possibly illegal goings-on lower in the ranks. In HBO’s Succession, we see the CEO Logan Roy repeatedly practicing what Matthias Gross has called “negative knowledge”—the knowledge to know which lines of thought, or information, would be unproductive or counterproductive to pursue. When subordinates attempt to inform him of a sexual assault and harassment scandal in his company’s cruise division, he repeatedly admonishes them: “Don’t tell me about that.” “My dad’s favorite employees are the ones who eat shit for him without ever making him aware,” son Kendall tells the head of cruises department when he tries to pass word of the scandal up. This is a clear strategy for avoiding later liability (which need not be legally enforced—social accountability works similar).

Maintaining actual ignorance to the effects or motivations of one’s actions can also be a strategy for removing one’s own internalized sense of moral responsibility—an evasion of the superego. Jacob Clifton, in his writings on Gossip Girl, writes that “[G]rowing up is a process of realizing that you don’t ever do anything by accident”—in other words, that one’s behavior is always strategic, even if these strategies are hidden from the conscious mind—and, through this realization, “getting yourself under control by figuring out the real reasons that motivate you.” As Bob Trivers has written about extensively, humans self-deceive so that they may better deceive others; in these cases, having an inflated sense of ego (i.e. being ignorant to one’s real skills and qualities), or having a somewhat massaged view of one’s historic accomplishments and actions, can be strategically advantageous.[5] It is likely that much of emotionality—anger, tears, hysteria, anxiety—is designed evolutionarily to be performative; these displays have clear effects on interaction. Anger can lead to individuals’ grievances being taken more seriously, and to deals being re-negotiated to avoid escalation of conflict. Tears can quick de-escalate anger and accusation. Hysteria and anxiety call out for palliation and address. Goffman asks: “When an individual supports a promise or threat with a convincing display of emotional expression, are we to believe him? When an individual seems carried away by feeling, is he intentionally acting this way in order to create an effect?” We might further ask: does the emotional individual himself know? Might knowledge of his own acting—and the self-consciousness, moral scruples, or performance anxiety this can entail—actually hinder his ability to act?

We will talk more shortly about threats, and what makes threats credible, but we can note quickly that if individuals are perceived as rational, then threats which go against the threatening party’s self-interest are unlikely to believed. But, if the threatening party can credibly indicate that he is ignorant to the full consequences of his actions—e.g. Schelling’s wildcard dictator, whose illegibility, impulsivity, and lack of attention to consequences allows him to more credibly threaten nuclear escalation than a dictator with a careful eye to self-preservation.

Knowledge also implies responsibility insofar as one cannot be expected to, and strategically would be better off not in charge of, some task at hand. There are “all sorts of things you might be pressured by others to do, which you can excuse yourself from doing if you make sure you don’t know how,” Sarah Constantin writes—”Witness all the people who are ‘just hopeless’ at housework or administration.”

Finally, it is the case that people give different accounts of the world depending on whether they believe observing parties are knowledgeable about the subject of the account. If one is explaining a theory, or a body of research, to one who is ignorant on the subject, he is far more likely to distort facts and present a strategically superior, which benefits him in some way, but less factual account, because he believes he cannot be held accountable, or his deceptions, exaggerations, and other rhetorical maneuvers will not be detected. At the same time, there are genuine epistemic advantages to “playing dumb,” insofar as it frequently elicits a pedagogical overview, or transfer of information, from the supposedly knowledgeable to supposedly ignorant party. My colleague Crispy Chicken has called this “Deliberate Beginner’s Mindset,” perceiving it as an attempt to explicitly reveal and explore, rather than assume implicitly, different groundings. To play dumb or ignorant, then, is to allow for either greater informational disclosures, or allow someone to reveal their true colors, motivations, or integrity—to let them “hang themselves.”

In other situations, lack of (the appearance of) knowledge is a clear liability. Certain kinds of knowledge can act as shibboleths, and we use our models of others’ knowledge constantly as metonyms for their intelligence, wherewithal, and socio-cultural background. Erving Goffman in his essay on expression games writes extensively about the subtle give-aways, in espionage, that risk revealing an embedded spy as a foreigner. In addition to more formal means of identity authentication, such as “identity tags” (usually state documents like a passport), there are informal means: 

…mental record of biographically relevant facts, for example, names of sibs, past employers, towns of residence, schools attended, regiments fought in, and so forth. More informal still is the run of information—including local geographical lore—which any resident of a claimed domicile is likely to possess. In brief, local cognitive orientation is required.

Goffman

This axis of identity interrogation is familiar to any who have been newcomers in, or attempted to infiltrate, a subculture and been asked their opinions on bands, artists, athletes, etc. “Fashionability… acts a passphrase, a shift key, a phase shift, a valuable proxy for speaker identity which then allows the speaker to communicate complexly, reflexively, with reference to self and modified by self. It is reliable because of the intense difficulty of faking fashion, which requires so much insider knowledge that any successful impostor is arguably no longer a fake. There’s a reason it’s tough to get into Berghain.” (src) And any minor slip can “out” an imposter, as displayed famously in the “three fingers” tavern scene of Tarantino’s Inglourious Basterds.

Knowledge or knowingness can be conveyed either explicitly, through verbal exchange, or implicitly, by making knowing decisions. “Beliefs leak,” Perry writes. “They leak out from conscious (and even unconscious) knowledge into behavior, emotional display, display of cognition, and ritual performance.” An individual who sets out for the bank in the wrong direction demonstrates that he is unaware of either his bearings or the location of the bank. This leakage, of course, is what makes the “reading” process of interaction possible—if players could perfectly control their impression, there would be no point in interpreting what is surely pure strategic manipulation. (That is, the leakage of true information is exactly what subsidizes misleading information, which postures as—mimics—the leaks.) But instead, there are constant pieces of physical evidence left behind by one’s actions and attitudes, which are expensive and sometimes even impossible to fully cover up. 

Commitment vs. Empowerment

We can start with a classic example of playing “Chicken”. Imagine you are driving towards another car, betting on who’s going to turn away first. If you throw your steering wheel out the window or spray paint over your windshield in advance, there is no way to steer or react in time, respectively. You have self-bound to a course of action. Your opponent no longer has a choice, if they don’t turn away, the two cars will crash. You know that your opponent is a rational agent, or at least rational enough to want to live, and you are engaged in a mixed game with quite a bit of overlapping interest (principally, survival). So this move is a winning one. Your opponent more or less is forced to veer out of the way, because they know for a fact that you won’t. It’s better for them to lose face then run into you and die. You’ve established a single Schelling point, a single outcome the system will coordinate to if no further communication or changes can take place.

In other words, I create this self-representation, I broadcast it, and then you optimize around a shared future reality that I’ve created, with respect to how much you believe my representation to begin with. Problems emerge because humans can represent one way and act another way. There’s nothing inherently binding about language. 

When you throw your steering wheel out the window, you’ve self-bound, which exists on a continuum with a costly signal: it is a way of communicating your capacities and thereby your future actions. It is a signal insofar as it still works by communication, it is still opticratic: what matters is appearances and appearances only. If your opponent does not know or believe you have self-bound, even if you have, the gambit fails. It is less that you have put a physical limitation on yourself, so much as you have communicated a physical limitation on yourself—that you can’t make a different decision anymore. You’ve said to him: I win or bust.

You are making a tradeoff, sacrificing a desirable action (veering off the road to save your own life, should the opponent call your bluff) to better secure the most optimal future: the one where I win and don’t die. You are “doubling down”; putting “all your eggs in one basket,” all your resources into achieving the optimal outcome. But problems emerge: you can represent things one way,  but act in another. Consider the theatrical effect of throwing out a dummy steering wheel and making sure the real one wasn’t visible through a heavily tinted windshield. This can lead your opponent to believe you are bluffing with your commitment; those who honestly self-bind are equally hurt by those who feign self-binding as those who feign self-binding.

Strategic ignorance can, in some cases, function as strategic commitment insofar as one’s options are ostensibly limited by one’s ostensible knowledge. Ethnomethodologist Ken Liberman calls this “performing obliviousness.” (We ought to think of ethnomethods, roughly, as patterns of strategic behavior individuals employ to legibly enter coordinating arrangements.) One pedestrian might look off to the side, or down at his phone, and walk straight forward in a way that at least gives the appearance that he is oblivious to the potential collision as he crosses a road. By broadcasting a state of ignorance (even if his ears are perked up, and he is looking out of his peripheral view…) he also communicates that he cannot get out of the way, and thus the cars must. His future state becomes as if a fact of nature, which forces others to adjust around him, similar to the tossed-out steering wheel. Lack of physical capacity (as in the game of chicken, where a driver tosses out their steering wheel) and lack of knowledge (the ability to know that one will collide with another pedestrian, and prevent it) function similarly in binding an individual to a course of action. And just as one can fake functional incapacity, one can fake ignorance.

Noncommitment as Empowerment (“Keeping Upwind”)

Intrinsic empowerment is the idea that, all else equal, it is best to maximize for having more degrees of freedom than less. If one does not have a particular goal in mind, but wishes to optimize, one ought to empower oneself by seeking capital which will open many doors, one when decides how one wants to use it. Paul Graham calls it keeping upwind:

Flying a glider is a good metaphor here. Because a glider doesn’t have an engine, you can’t fly into the wind without losing a lot of altitude. If you let yourself get far downwind of good places to land, your options narrow uncomfortably. As a rule you want to stay upwind. So I propose that as a replacement for “don’t give up on your dreams.” Stay upwind.

The downside to self-binding, of course, is that it is precisely the opposite of keeping upwind. To move in any direction toward a specific goal, while necessary if one wants to accomplish a goal, lower one’s degrees of freedom insofar as one is now further away from goals which lie in the opposite direction. In keeping upwind, agent “intrinsically” empowers itself at the cost of foregoing situationally empowering itself.

What are the benefits, however, not so much of being upwind but of appearing upwind? Primarily that one’s opponent cannot optimize around you in turn. If you have committed to an invasion at a certain cross-section of coast, and have moved major parts of your naval force there, you are now disempowered attacking anywhere else, and the defending army can more safely concentrate its resource at that cross-section in turn, matching your situational gain in force (by committing) with a corresponding accommodation of their own. Always, they are playing the percentages.

Footnotes

[1]

Pat Symonds, a Formula 1 CTO, discussing his team’s use of computer models for F1 Magazine: “The primary method we use is one known as a Monte Carlo simulation. This is a long-established statistical technique whereby a large number of Virtual races are run, each with different parameters, and pitstop laps are applied to all the cars. The many thousands of results are then analyzed to determine the probability of a given outcome for a particular set of decisions. If we relied on this alone, it is likely that everyone would come up with similar answers, so we also apply a technique known as game theory. This covers many mathematical techniques but, for example, takes the knowledge that people will deviate from a deterministic optimum to take advantage of the undercut. With this assumption we modify our tactics to try to counter their move.”

In an undercut, a trailing driver makes a pitstop earlier than the car ahead of it; this gives the trailing car fresher tires, and allows them to drive the next few laps faster than the leading car, which has well-worn tires. When the leading car is inevitably forced to take a pit stop itself, the trailing car has now gained enough distance, with its fresh tires, to “jump” the lead car (pit stop times being equal). While the once-leading, now-trailing car has fresher tires than the undercutting car, depending on where in the race the undercut is performed, the undercut’s jump can provide an overall advantage in finishing order.

[2]  The picture is far more complicated in reality, and racing is far from being a game of pure conflict. Players share an interest in avoiding accidents, as noted, but additionally drafting, as well as NASCAR’s point system, encourage provisional as well as long-term cooperation among drivers. In long drafting lines, a given driver embedded in such a line will risk losing serious momentum if he breaks out alone, and he will look for information that the car ahead or behind him will break off with him—while also wary of a betrayal in which either driver signals a break, then defects, leaving him stranded outside the draft line. Radio communication is often used by spotting teams to arrange provisional agreements, and since (like police radio) this communication occurs on open frequencies, it can be monitored by rival teams. (Ronfeldt 2000, “Social Science at 190MPH”)

[3] Lynch’s film work, which is paradigmatically creepy, centers thematically on the dark, seedy underbelly of otherwise idealized locales: white picket fence suburbs, sunny Los Angeles, small towns in the Pacific Northwest.

[4] One prominent exception is something like art or magic, where audiences are actively interested in being fooled. The difference likely is that these are “sandbox” domains, without real threats, where false beliefs hold no serious consequences.

[5] Trivers’ idea is that there is an evolutionary arms race between humans’ ability to lie and their ability to detect lies. Trivers’ proposal is that this arms race would eventually result in much of our knowledge or beliefs “going underground,” into our subconscious. Without conscious awareness, our conscious self strategically kept ignorant by our subconscious, we can more confidently and robustly bluff: we truly believe, at least at one level.