This essay is the first in a ten-installment series, but can be read as a standalone introduction:

1. All Communication is Manipulation
2. All Communication is Behavioral Manipulation
3. Consensual and Nonconsensual Manipulation
4. ACiM is a Natural Extension of Cybernetic Theory
5. Response to Simpolism on ACiM
6. A Landscape of Communication
7. All Is Well E28: Functional Railroading
8. All Is Well E28: “My Hands Are Tied”
9. Girls S2: Lip Service
10. Against Self Expression

---

All communication is manipulation. Some manipulation is mutually advantageous.

Alternatively phrased, the purpose of communication, broadly, is the alteration of others’ actions. Or, Communication is defined by its interest in altering the receiver’s actions. Actions, expressions, and speech that do not attempt to manipulate receivers’ behavior are not, properly considered, communication.

It’s clear how this is true of imperative statements: we give orders so that they will be followed; we make requests so they will be fulfilled. But it is equally true of indicative—that is, representational—expressions. We alter others’ behavior through our strategic representation of reality, including but not limited to our own internal state. These representations carry transformational potential precisely because receivers—all of us—carry a crowd-sourced model of reality, on which we base our actions. Alter this model, alter behavior.

(This crowd-sourcing, whereby we allow others to tell us about the world, and tell us about themselves—as the predator allows the dart frog to tell it about itself, through its red stripes—provides us with knowledge far beyond our experiential purview. In cooperative arrangements, agents can multiply one another’s knowledge base many times over, a process that has so far culminated in written text and cultural transmission. But this crowd-sourcing also opens us up to sabotage—the spoiling of our agendas and aims through bad information.)

To be more precise, imperative statements are nothing more than a category of indicative (representational) statements, a class that specifically represents the desires of the requesting individual to a receiving party who carries an interest in fulfilling the speaker’s desires. This interest may be on account of a formal, institutional incentive structure, or because the receiver believes that the speaker’s desires and outcomes are aligned with his own, or because the receiver wishes to ingratiate himself with the speaker, or because they are engaged in a relationship of reciprocity, etc.

The basic idea here is excruciatingly simple. Knowledge and desire together determine our actions, what we might call “tactics.” The reality (of the environment, of our own abilities) determine the outcome of our actions. Such actions co-create the environment in which our actions takes place, and co-create the state of knowledge which inform actions. This is the state of “ecological huddle,” in which each of us has a stake in the actions taken by others, and in which each of us has an influence over the actions taken by others, through our representation of reality on which they base their actions. Thereby we represent to alter actions to advance our goals.

II.

There are a series of objections which come up when I make this claim about the purpose of representation (and by extension, communication, which I see as the process of creating and interpreting representations). The rest of the post will use these objections to flesh out exactly what this thesis—that all communication is manipulation—entails and does not entail.

Those objections are:

• That I am using the word “manipulation” in a non-standard way.
• That this thesis implies a highly adversarial view of human relationships and interaction.
• That much of everyday conversation and interaction is undertaken in an open-ended spirit, that is, without particular goals or outcomes in mind.
• That the identification of all communication with manipulation erases an important distinction in our everyday language that is usually upheld by the word “manipulation,” where some modes of deceptive, abusive, or adversarial interaction are distinguished from more pro-social, honest, cooperative modes.
• That this view of communication disenchants the social.

To be clear on my meaning of the word “manipulation,” I mean it not in the pejorative, colloquial sense of manipulative behavior, but in the more neutral sense we refer to manipulating objects—the alteration of state or behavior, nothing more. It is a morally neutral descriptor.

But before I address these (explicitly advanced) objections at length, I want to address what I see as an implicit claim underneath all these claims, which is that this view of communication is a novel perspective on interaction that stems from an evolutionary psychology view of the world. It is true that this “rationalizing” view of human interaction as motivated has been identified, in the past, more with economics, evolutionary theory, and war gaming than with humanistic traditions. (And in this, there is a real sense in which such views are disenchanting; they make explicit the rules of the game, rather than supposing the often romantic, idealized narratives surrounding motive which game-players attempt to create and sustain.) But the idea that all communication is manipulation is not my own invention; it is a compression of many thinkers, emerging from diverse disciplines, who have argued something similar.

Richard Rorty, in his essay “Freud and Moral Reflection,” identifies this view—”that everything everybody does to everyone else (even those they love blindly and helplessly) can be described, for therapeutic or other purposes, as manipulation”—with the Freudian school. Karen Pryor, coming out of a marine biology background, writes in Don’t Shoot The Dog that, “to people schooled in the humanistic tradition, the manipulation of human behavior by some sort of conscious technique seems incorrigibly wicked, in spite of the obvious fact that we all go around trying to manipulate one another’s behavior all the time, by whatever means come to hand.” Adam Smith, in his Theory of Moral Sentiments, writes: “The offering of a shilling, which to us appears to have so plain and simple a meaning, is in reality offering an argument to persuade someone to do so and so as it is for his interest. And in this manner every one is practicing oratory on others through the whole of his life.”

For my own part, I became convinced of this definition of communication from reading ethological accounts of animal signaling, where communication is defined quite simply as occurring whenever “one individual’s actions provide a signal that changes the behaviour of another individual.” Setting aside for the moment human behavior, it is the uncontroversial case that, in all studied superorganisms—in body cells, bacteria, animals, and at the extra-human level of institutions and organizations—the function of communicating, to other agents outside oneself, is always the manipulation of those agents’ behavior, in a direction that advances the goals of the speaker. The company produces public relations material and advertising. The bird puffs out its plumage or sings its song to attract mates and intimidate rivals. The symbiotic bacteria fire off signals to one another, “so that” they may synchronize their behavior. We may err in our teleological language here; there is no intent in the way we usually think of intent, but there is a fact of the matter, which is that if cell A fires signal X, cell B will perform action Y, and that this relationship of informing or manipulating is evolutionarily advantageous and stable. Communicative actions, or signals, which did not improve the (reproductive) prospects of the communicator, disappeared through natural selection. Dominance, survival, strong alliances, securing food or other resources, are all means or surrogates to this end, and thus speech which secures them is preserved.

(We can think of intelligence, here, as the ability—discovered and honed by natural selection—of an entity to be altered by information. A difference that makes a difference, that alters agent behavior—this being selected for by evolution precisely because dynamic, anticipatory response by an organism to its environment is advantageous.)

Since a signal cannot move rocks, or bring rain, or act in any way on non-agentic matter, it must improve the signaler’s situation by altering the behavior of those who receive the signal.

The receiver is a medium, a vessel for the physical changes which eventually result from the information received (sometimes after a chain of further transmissions). In the highly symbolic world of human social reality, this chain can be long and convoluted. When an upstart politician airs a campaign ad, he does so that we will vote for him, so that when the information is tallied, and he is declared winner, the incumbent will step aside, and those authorities vested with the symbolic power to consecrate the transition of power will consecrate it, and individuals (from everyday citizens to bureaucrats and police enforcers) will therefore follow his decrees through physical acts, etc. He has communicated in such a way as to cause further chains of communication, which downstream will alter the physically relevant actions of empowered individuals. No matter how long and complex these chains of altered communication are, they must always bottom out in the alteration of physical action—the wielding or withdrawal of force.

III.

Given this apparent fact about the role of communication in non-human entities, the burden of proof shifts from proving that human communication is manipulative, to proving that it isn’t.

The most common argument against this view of communication tends to be founded in introspection. This argument goes that the objecting individual does not feel like he is trying to alter the receiver’s actions, that in fact he feels quite the opposite way—his approach to communication is “open-ended”; he would be upset to learn that his conversant felt “railroaded” by the interaction.

If I can risk extrapolating from my own phenomenology, and the introspections of those who have advanced this objection, I concede that it does not, often, feel as if we are trying to manipulate others. We often speak for the pleasure of it, or because the words feel appropriate. And yet, when we study bees, and watch them dance, and after watching the dance, other bees fly to a location previously scouted by the dancer, we say “the purpose of the dance is to communicate a location for a new hive,” which is to say, “the purpose of the dance is to move the colony to the scouted location.” The phenomenology of the bee is more or less irrelevant here; it may well dance, and dance the way it does, for the pleasure of it.

IV.

What we are left with, here, I think, is that there are two very different perspectives on a system. One is functionalist, the other psychological. One looks only at the patterns of behavior, and the patterns of effects which stem from this behavior, and defines the behavior in terms of its effects. The other is interested in the goal which the individual actor consciously believes himself to be seeking out, his own self-ascription of motivation as it feels from the inside. To be clear, the description of communication as manipulation is a functionalist description, not a psychological one.

What is crucial is that these two motivations are more or less synchronized, which is to say that the descriptions are more or less compatible. The actor may act on the basis of a feeling, or a desire for pleasure, when he initiates sex with a person of the opposite sex. He may say it was “for” pleasure, and in a sense it is. And in another sense, it is clear that this behavior is about and for reproduction.

I think the best way to square this is with the concept of mesa- and base-optimizers. A base-optimizer consists of a selection algorithm which determines which embedded agents survive. The agents it selects between are termed mesa-optimizers; they have their own priorities, but their priorities have been selected and trained against the logic of their base-optimizer, such that the two are in rough, imperfect alignment. Natural selection is one such base-optimizer: organisms who feel pleasure upon sexual intercourse end up reproducing—are selected by the base algorithm—even if the thing they are optimizing for, in their place as mesa-, is pleasure. Divergences in goals—misalignments—emerge whenever the environment changes the careful, contextual coupling between mesa- and base-optimizer goals: birth control is one such major uncoupling, such that pleasure and reproduction no longer go hand-in-hand. As a result, birth rates have plummeted since their introduction.

Similarly, within a company structure, a manager who hires, promotes, demotes, and fires employees is acting as base optimizer for the mesa-optimizers that are his employees. Let us say, to simplify, that his selections are based on how the employees advance the company interest. Employees who show up for work early in the morning, bleary-eyed, or who turn in their inventory reports on time, do not typically do it “for” the benefit of the company. They do it so that they are not reprimanded or eventually fired; they do it to maintain good standing in the training system or selection algorithm of the base optimizer. And yet they clearly also do it “for” the company! The objective outcome of their behavioral patterns is, broadly, to advance company interests. Even as the internal, psychological logic is one of personal incentive and reward, the external, functional logic is defined by company interest.

(This is complicated by the fact that institutions are organized with a nested hierarchy of base- and mesa-optimizers running selection games on the mesa-level below them, e.g. the company manager himself is selected by a higher-up; the company stock lives or dies on public purchase; etc. But we can set aside these complications for now; they have been handled elsewhere.)

And so, in conversation, when we speak and optimize and act “so that” others have a good impression of us, this good impression may feel like all we care about as individuals, but it is also clear that creating a good impression (and creating a very contextually specific kind of good impression) will improve our situation, lead to better outcomes in most cases than a bad impression. Or, when we speak “in order that” we get rid of some nagging feeling which is eating at us, what is this feeling, if not a surrogate, a proxy, which has evolved to yield outcomes which broadly advance our interests. Evolved not just biologically, over generations, but culturally, through the training system that is society. We make millions of utterances and representational moves in our lifetimes, from the clothes we choose, to our facial expressions, to the way we talk and write. And though causality is complex, in a welter of many simultaneous, bundled communication moves or “tactics,” we do note, and feel strongly, the effects of a strategic misstep: when we turn someone against us without desiring to; when we upset someone we wished not to upset; when an interaction goes south and does not end how we hoped it might.

V.

Now that we’ve addressed the general insufficiency of introspection in understanding the functional purpose of communication, we can address the specific claim, and disprove it even on introspective grounds. To recap, this specific claim is that speakers do not hold specific, narrow agendas for the behavior they want to evoke from receivers. They may just want to “have an honest conversation,” to get their interlocutors to “speak their mind” or enjoy themselves or say whatever feels appropriate. Even from a functional perspective, this can well be true—there are always certain axes on which our desire, for an outcome, is underspecified or left vague. We may hope for a certain tone, while caring nothing of the content, or vice-versa. In a police interrogation, the defendant’s tone, during a confession, or the order in which he tells the story, may be of such low priority to the interrogator’s goals that he is functionally agnostic. Rather than seek incredibly specific outcomes, we recognize that there are entire classes of outcomes which are approximately commensurate in terms of advancing our immediate and long-term interests, and we are more or less agnostic to which entry from that class ends is selected by the interlocutor. What matters is the general class.

Similarly, we can note that, if we were trying to have an “open-ended” conversation about psychology at a party, and one interlocutor keeps answering our comments or inquiries with criticisms of the event catering, we may quickly get frustrated and exit the conversation. So much for “agenda-less” interaction. Clearly we wanted something, wanted our interlocutor to behave and act in a relatively specific way, even if it this specific way leaves many implementation details up for grabs. We learn here via negativa—learn our hopes through their upset and our resulting disappointment, learn our expectations from our reactions when they fail to materialize.

To summarize the arguments so far:

• One: that we ought not give much credence to our phenomenology or introspection in determining what motivates us from the functionalist perspective which the phrase “All communication is manipulation” takes. Rather, we should look at objective patterns of behaviors and effects, since these effects are the thing on which we are culturally (in our own lifetime) and biologically (over many lifetimes) trained—the reinforcement loops which select for and select out.
• Two: that our default assumption ought to be that human communication serves the purpose of altering receiver actions, since this is true of all non-human communication, and we are nothing if not bound up in the same environmental context, the same strategic situation of conflicting desires, and the same family lineage, as non-human organisms.
• Three: that there are different levels of rationale behind our behavior, and that many of the things we desire, we desire instrumentally insofar as their attainment tends to correlate with other valuable outcomes, for instance, we may eat “for” flavor but our taste buds have evolved for nutrition, thus we eat “for” nutrition.
• Four: that our aimed-for outcomes, in manipulating others’ models of the world via representational work, are always both narrow and broad, specific and vague. The set of desired outcomes is a very small sub-space of all possible outcomes, and yet still leaves serious optionality and degrees of freedom on the part of the respondent.

VI.

To expand on the first point: there is a large body of literature, from psychoanalysis to evolutionary psychology, that distrusts self-reflection, in terms of feelings and phenomenology, as a source of behavioral insight. These theories broadly hold that we self-deceive in significant ways, and especially that we self-deceive about our motivations. That we do this either to prevent the psychic pain (or dissonance) of acknowledging some fact about our motivations, or in order to be more strategically competitive in the social arena, since by deceiving ourselves we are better able to deceive to others. Which is to say, it is no coincidence that manipulation is both the air we breathe, socially, and that it is deeply stigmatized, socially. We are wary and terrified of being manipulated even as we are nothing but manipulated. As soon as we learn that someone has an agenda or is motivated to speak in the way they do, to represent reality the way they do, we distrust them. And yet all our representations are motivated, why should we speak without purpose, why would we act when it changes nothing, or when it would leave the world no better than if we did not speak, or when it would leave the world worse.

(This norm of suppressing the manipulative quality of communication extends even to protecting those who manipulate us, and not merely in cloaking our own strategic representation-work. An individual who advances a highly biased interpretation of a situation will see his account taken with a grain of salt, but the receiving party is more likely to call the matter “subjective” than he is to accuse the account-giver of manipulation. An individual who advances objectively disprovable claims will often be considered merely mistaken. It is only when it is abundantly clear that the individual [1] advances demonstrably false claims [2] knew very well that these claims were false at the time of advancement, that we will accuse him of attempting to manipulate us by presenting an inaccurate portrait of the situation.)

And yet as a result of this stigma our agendas must go underground, even as we have become deeply confused about the behavior which we fear. That is, we do not manipulation in the neutral sense, but a dishonest manipulation in which we are led, under false guises, to act in ways which hurt our own best interests. And yet most representation is in some way dishonest, in some way elides true facts which are unbecoming to the speaker. As Goffman notes, members of the clergy, asked why they entered the profession, never mention that it helps advance their social standing—although this is almost certainly a strong motivation. (How does Goffman guess they are hiding this motivation? Because it is a clear, objective, externally visible benefit of the clergy, and Goffman assumes that clear benefits of some action become attractors for that action, even if it does not “feel” like a or the reason for the actor.) And when we argue for something, we rarely acknowledge to a full extent the merit of the other side; even when we try to, it is likely our brains do not put quite as much effort into investigating the merit of the other side—is this not manipulation, then, to argue? Is this not dishonest representation?

The line between what we find repellent, in terms of manipulation, and what we expect or find banal, remains unclear to me. Certainly, honesty and outcome matter, but they are nothing near sufficient and necessary criteria. Likely, as with most of our language, the actual way we use “manipulation” in the social, pejorative sense does not have some clear conceptual logic, but rather is dictated by a pragmatic logic: we call things manipulative as a tool, a gambit, to either change the behavior of the individual we find manipulative, or to change other’s impressions and beliefs about the manipulative individual, to lessen the base of his support, and perhaps extract revenge or punish him. The very blurriness and ubiquity of manipulation serves to make the term a weapon without check—almost any speech or behavior can be cast, accurately, as manipulative, since it is manipulative in the strict sense we’ve used throughout. And yet this sense of manipulative, of attempting to alter an individual’s behavior, is not, cannot be, the criteria by which we judge actions to be moral trespasses.

We have reached, then, the objection that defining communication as manipulation, and defining manipulation as the morally neutral alteration of an agent’s state or behavior, subverts a critical distinction that the term “manipulation” presently connotes—as distinguishing abusive, coercive, or deceptive behavior from honest representational work. This might be true were I word-tzar, abolishing overnight this distinction, and leaving only a void of language. But we are students, far from influencing the mainstream of thought, and by acknowledging that the difference between abusive, coercive communication, and honest or cooperative communication, is not a difference in whether the speaker is or is not attempting to alter receiver behavior, then we can go looking for where, in actuality, that difference lies. We are forced to identify the actual communicative behavior we find reprehensible, which are not and cannot be “the alteration of the receiver’s actions through strategic representation.”

The situation is similar to the use of “selfishness” in evolutionary psychology. John Nerst writes in a post at Everything Studies:

Alright, what’s bothering me [about Simler’s essay “Leaning Tower of Morality”]? The conflation of “self-interest” in an evolutionary sense with self-interest in an ordinary sense.

When we call somebody “selfish” we have certain things in mind: screwing over others, not cooperating, being explicitly calculating in order to only maximize one’s own gain in interactions with others, hurting people for personal benefit, being untrustworthy, ignoring costs to others etc. We use the word to classify actions and those who perform them, and specifically to indicate disapproval.

That’s what we care about when we fear and dislike selfishness, and “genetic self-interest” is not this at all. It’s a completely separate thing, with an altogether different subject (genes, not people). We can and do have altruistic instincts and knowing how evolution created those won’t make them any less altruistic in any way we care about.

My position is similar to that of Simler, who responds to Nerst’s complaint:

The most important distinction between morally good people and morally bad people is whether they’re playing the sociopath strategy (exploitative, negative sum) or the prosocial strategy (cooperative, positive sum). Of course it’s probably more of a spectrum than a dichotomy, but that’s the main axis. Both strategies are genetically selfish, but that’s irrelevant: we still judge the sociopath as evil and the helpful neighbor as good, and we’re right to do so.

VII.

To expand on the second point—that human communication as manipulation ought to be our default assumption, given it is true of all non-human signaling—it is also worth nothing that, as communication is a class of action, and we broadly expect for our actions to “have a purpose” or reason—that is, to alter the world in a way which brings it more in line with our ideal, an ideal which can be selfish or selfless—then we should expect communication to behave the same way. That it is motivated, that it desires to have effects, and—since the only way information can have effects is as interpreted by another agent, thus altering that agent’s representation of the world, and thereby as a final step, altering his actions so as to eventually achieve these effects. In other words, once we understand the narrow way communication (expression, representation, speech, etc) can have effects on the world—by using another agent as medium, and only by using another agent as medium—this view of communication as manipulation flows naturally from our understanding of communication as a class of action.

VIII.

We have dealt with the objection that much of human interaction is undertaken in an open-ended, goal-less spirit, the argument that I am using “manipulation” in a non-standard way, and by extension, that my non-standard usage erases the pragmatically useful distinction which the term “manipulation,” in its current, pejorative form serves. Two objections remain: my definition of communication implies an adversarial view of human interaction, and that this definition disenchants the social.

Hopefully, in clarifying what I do and do not mean by manipulation, I have foreclosed the former objection. But it will be useful to concretely show examples of cooperative or pro-social manipulation in order to really demonstrate how ethically neutral strategic representation is, at its heart.

There are two broad ways a strategic representation can be considered cooperative or pro-social. First, it can strategically create an impression or model of reality, in the receiver, which is as accurate as possible, within the bounds of relevance, to what the representing party believes to be the case. We call this mode “honest,” and note that it does, indeed, take quite a bit of strategy to accomplish: not just the ability to compress noisy data to an accurate signal, but the theory of mind to present this signal in a semantic form the receiver will understand as intended. The second way a representation can be strategically pro-social is by presenting a disbelieved account of reality, or instilling in the receiver a model of reality which is discordant with the speaker’s own model, in an attempt to improve the receiver’s situation. We call this mode a “noble lie.”

In many but not all cases, the manipulation will be transparent, at least at one level of aim. Following the pattern of Gricean linguistics, an expression will be conveyed in the form of “I intend you to form a belief based on my speech act, but I also intend you to recognize that I have that intention.” (Actions and beliefs being roughly symmetrical; altering beliefs matters because beliefs cash out to actions.) This transparency is of course limited in scope: When I share a story that conveys a given moral, you may recognize that I wish to convey said moral, without realizing that the story is false. That is, we often understand the effects others desire even as we are being deceived: we understand that their speech has been designed to accomplish the effect it accomplishes even as we do not know whether the representation accomplishing said effect corresponds to reality (or the speaker’s picture of reality).

When we are asked a question, we understand the speaker desires an answer. When our name is shouted, we understand the speaker wants our attention. We recognize people “go fishing” for compliments or praise, for sets of effects; often, we readily give the desired responses out (although some personalities seem to grate against such implicit requests, and refuse to give people what they’re after).

Sometimes, we are perfectly transparent about both our desire to be manipulated, and our manipulation of others. These cases include requests in the business world to “pitch me” or “persuade me,” or dating dynamics in which (archetypally, a woman) desires to be talked into a relationship, or into sexual engagement. When we feel torn or ambivalent, we may call upon others to resolve our ambivalence through a strong rhetorical showing, and will feel swayed to one position or another even as we acknowledge that the account we have received is highly partial and highly interested (in the sense of motivated or biased).

And as Amirism has noted on Twitter, it is often only when our manipulations fail that others identify them as such. When we try to cheer someone up, or make them feel validated, we are obviously manipulating them—but in many cases, they will (happily) go along with it, glad for the alteration of mood, unless of course they are not in the mood for alteration, in which case they might snap, “Stop trying to control me.”

(We can add “cheer up” to our list of uncontroversial manipulations—alongside attempts to excite, encourage, intrigue, entertain, impress, charm, galvanize, scold, humiliate, shame, and all sorts of affective alterations, whose ostensible justification is typically behavioral.)

When we “enable” someone to do something which he otherwise would be unable to, this is not viewed as manipulation, because our implicit view of free will and conscious choice do not cast the enabling as “causing” the behavior (merely as “allowing” or “assisting” it). But taking a functionalist perspective, and brushing aside the illusory cloaks of choice and consciousness, we can see that the enabler has caused the resulting behavior, and that whether or not he anticipated the behavior would result, and whether or not he “enabled” in order to bring about the resulting behavior, the causal structure of interaction is the same.

Manipulation is simply the only way we can alter agents behavior informationally (as opposed to via physical force). And the proportion of informational manipulation to physical force manipulation is much higher than we might think: even when we beat a dog for begging, we do not physically prevent his begging in the future, rather, we create an informational association that alters his model of the world, such that the dog is averse to begging because he believes, as part of his world-model, that a beating will follow. True, purely physical incapacitation and alteration of one another is both rare and taboo in normal human social life; it becomes common only in times of social breakdown, during crime and wartime.

As with any other strategic dynamic, how cooperative or adversarial manipulation is is closely related to the capacities of agents for mutual monitoring, to how long- or short-term the interpersonal relationship is, and to the stance and proclivities of the speaker (which are non-trivially evolved and acculturated; we are arguably more honest, in modern life, than we need to be, but we have been conditioned by culture and evolution to be so honest). Long-term alliances and relationships, which are crucial to the mental health and strategic efficacy of most individuals, require that each participant produces (the impression of) comparative value to each other participant. In a sense, we could say that all manipulations must, in the very long-term, approach consent, since by severing the bond of the relationship, the individual frees himself of the manipulator’s effects. Crucially, the “official story” (the implicit social account of why a given representation was advanced) matters less than the representation’s real effects: an individual who gives underhanded compliments, or finds ways to make others feel and act poorly, against the larger macro-picture of how these individuals would like to feel and act, will find himself socially disconnected.

IX.

Whether this view disenchants the social, I don’t know. Whether information is known privately or publicly alters the game state quite a deal. If this view of communication were to become mainstream, it might change a great deal. On the other hand, even the most famous philosophical accounts of language from the 20th C remain unknown to virtually all outside the discipline.

When we come into some new theoretical frame, we may respond in very different ways, depending on its claims. Learning that the sky is not actually blue, that it is a trick of atmospheric dust, does not change our experience of the blueness of the sky. This view says: all communication is manipulation, but also, human communicative behavior is exactly as we already think it is. It does not imply more selfishness or more dishonesty in human conduct, it does not imply that we are Machiavellian optimizers. It merely reframes what we already believe about human communication, by conceptualizing it through the frames of outcomes, and pointing out that necessarily, outcomes of communication must use human beings as their medium.

The second way to take such a theory is more revisionary. Our motives are not what we thought. We are more outcome-motivated, and more performative than we like to think. We are constantly attempting to further goals and meet desires through our communications. We do not neutrally represent “the truth.” Our unconscious, expressed through words and gestures and choices of display, is constantly reaching out to manipulate others around us, even as our conscious is often innocent to these doings.

Most likely, the reality is a little bit of both. To believe in this theory will change your view of others’ speech and of your own. It may cause self-consciousness, for as soon as we become conscious of what we are up to, and own up to the strategic shape of our speech, we are now responsible for them. We must accept responsibility for their effects as we accept responsibility for any other tool we wield. The flipside of disenchantment is always denial, and while the costs of disenchantment are clear, the costs of denial are equally so.

X.

Tallying things up, where do we stand?

On the one hand, I think “all communication is manipulation” is structurally analogous to pragmatic or functionalist linguistics—the idea that meaning resides in expression’s effects. Extended to discourse, its natural result is torque epistemology.

In an internal doc about text transformers, recently circulated around the Inexact Sciences server, Crispy wrote: “We do not have tests for whether people understand us, we have tests for whether people fail to understand us, and if they don’t fail consistently, then we tend to have some confidence they understand us, at least in some aspect where they haven’t failed too much.” The question becomes: How do we ascertain that someone (or something) has failed to understand us? And the answer, I think, is that it responds in ways that do not seem to follow from our own expressions. Its utterances do not “build in” the expressions which preceded it. If we have asked it a question, it neither answers that question nor rebuffs the question’s validity. In other words, it does not seem to understand what we want from it.

In the server I wrote: Are people gonna hate me if I suggest that pragmatically, “It understands me” is shorthand for “I am able to manipulate it informationally”? Looking back, I think that’s not quite right. The other way we tell that something understands us is if it reads our intended manipulation, and rebuffs it.