This is the second installment in a ten-part series.

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

---

I want to establish, from the get-go, the uncontroversial, borderline tautological aspect of what I mean when I say “All communication is manipulation.” As a recap on what I mean by “manipulation,” I define the word as “the alteration of an agent’s behavior.” When we sum up these uncontroversial aspects, I think we’ll find that they persuade us of the controversial aspects as well.

The uncontroversial aspect is this: any utterance or expressive act, adequately displayed and registered in the interaction as a source of information, will alter the behavior of the receiver. @joXn mocks, on Twitter: “Did you realize that when you interact with someone in any way you literally change their sensory inputs? Makes you think.” This is indisputable; what is also indisputable, I think, is that any communicative act or “unit” (whatever that might be; perhaps a unit which is non-interchangeable, or which as a unit can be said to “have a point”) will in some way alter the mental model of the receiver. The receiver is automatically building (learning) a model of the interaction and the world, a model which incorporates the representations of the expressing or communicating individuals.

Moreover, any difference in mental model corresponds to a hypothetical shift in behavior given some hypothetical scenario which leans on that difference. By extension, any sufficiently large and sufficiently relevant (to the receiver) difference will change his actual behavior.

These can be both immediate, local alterations, which occur inside the interaction, and more long-term alterations which occur outside the interaction. When we respond to someone, our response typically incorporates the specifics of their previous utterances and self-expressions and representations more broadly. That is, their communicative acts have altered our actions. However, even when we leave interactions, we often find ourselves behaving differently as a result of information received within the interaction.

For the most part, I think these claims are uncontroversial. What becomes controversial is the implicit claim to essence or purpose that “to be” words drag with them: Communication is manipulation. Not the uncontroversial verb form: Communication does alter receivers. Moreover, if I say, “one facet or aspect of communication is its effects on receivers’ behavior,” I doubt this will be controversial either.

In a lot of cases, I am very wary of the “to be” form, of any discursive “X is Y” move, which I’ve called narrow-and-conquer tactics.

But in this case, I think that communication is only manipulation, that communication’s central purpose is and has only ever been manipulation, that there is no adequate way to understand communication that does not found itself on communication as manipulation, etc. This is a strong claim, and one I keep talking myself out of and into again. Today I am in an “into again” mood. The effects of communication are its evolutionary function, and they are its cultural function, and they are its case-by-case individual function, which reverse-engineered means, purpose. Effects can only occur in receiving agents, end of story. Communication is manipulation.

If we are OK saying that information is a difference which makes a difference (to quote Bateson)—if we are OK saying that non-differences, or differences which do not make a difference, are not information—I think we need to bite the bullet that by extension, communication—writing which is read—is defined by its “making a difference” for its reader. And a difference in reality which is irrelevant to future actions is not a difference worth updating one’s mental schema with.

Communication is not always intentional, but when it is intentional it is intended to make a difference. If it was not intended to make a difference, it would have no point. It would “mean” nothing at all.

Not all information that is expressed (“written”) is picked up by a reader. Not all information that is read was intentionally written for the purpose of being read (by that particular reader). But it was written nonetheless—perhaps as a byproduct of some other action, the way snowmobiles leave tracks behind them; or perhaps written for some other audience, the way a politician’s private comment becomes headline news. Whenever something that has been written is read, we can say that an act of communication has taken place. Any “bit” of information which alters the reader’s model—any difference that makes a difference—we might call communicated. I am not so much making an argument here as I am elaborating on a definition of what communication is.

If I reassure you, at the very least my aim is to alter your mental state. If I try to cheer you up, or get you down, to disparage you in public, to humiliate or shame you, to gas you up or entertain you, I am always attempting to alter the mental states of those observing my actions.

We will deal only with intentional writing for the rest of this post.* An intended writing is an action, and actions always have a point. A point or purpose is synonymous with effects, and the effects of information can only be accomplished via an altered observer, a reader altered by what he has read.

The next question that comes up is whether communication is more concerned with altering mental states, or more concerned with altering behavioral programs. We can call the former “affective manipulation” and the latter “behavioral manipulation.”

It would, I think, be uncontroversial to say that behavioral manipulation is the central purpose of communication in group coordination activities, for instance, when a quartermaster on a ship gives a command to a ship’s mate, it is expressly with the purpose of that mate executing the command, i.e. performing some resulting action. No one, not the sailor nor the quartermaster, would dispute this is the function of the communication at hand.

Nor would it, I hope, be controversial to say that, in conflict—such as a high-stakes poker game, between competent players—the purpose of intentional communication is the manipulation of other players’ tactics or choices. This is why Schelling recommends that players, in such adversarial situations, may wish to “blind” or “deafen” themselves, so they cannot receive any such sensory inputs from their adversaries.

The question now arises whether these group coordination, and individual or inter-group conflict, scenarios are central archetypal of communication. I believe they are, since from an evolutionary perspective, they are undoubtedly the original function of language, as evidenced by the purpose of signaling in non-human organisms. However, in modern life, casual sociality—social or familial phone calls, small talk, online chit-chat, etc—occupy a greater amount of many people’s lives than group activities or inter-agent conflict. Participants in casual sociality would not typically ascribe to themselves the desire to influence or alter other players’ actions, as they would in the quartermaster or poker examples. The phenomenological explanation no longer supports the functional explanation, and we are stuck choosing an epistemic authority to resolve the matter.

I think there are a few reasons we should be wary of prioritizing the mental over the behavioral even in the realm of casual sociality.

One is that we only ever interact with the behavioral. We build up models of others’ mental states, to be sure, but only through their expressive behavior, which is to say, through their communication of mental state. If writing—expressive representation—is an action, then we only have actions. We create actions, and we receive actions; the rest is speculation, inference whose basis is read actions. When we try to assuage depression or jealousy or anger, we can only ever diagnose via symptoms, and we can only ever know if we have assuaged via the disappearance of symptoms. So from the perspective of a control system, we act on behavior, and we respond to behavior, end of story.

When we learn effective interventions, which we take into future interactions, we have two sources: the real and observed consequences of previous actions, and our own personal experience of mental states in response to stimuli. Of these, I think anyone with competent theory of mind learns to prioritize the latter as authoritative. Even if, from personal experience, we believe an utterance ought to have positive mental alterations on a receiver, if the receiver consistently reacts badly, we will almost certainly update our interactive model. In other words, actions, and not our own phenomenological projections, are the authorities in guiding our behavior.

Moreover, it is the behavioral alterations which have real consequences for us. While we may be hypothetically concerned with the mental states of others, our own mental states, of which we tend to be centrally concerned, is the direct, knowable result of others’ actions (and only indirectly, unknowably, the result of others’ mental states). We are trained over our lifetimes to make certain communicative moves and not others, and this training is the result of real consequences, not hypothetical and impossible-to-access mental states.

Finally, it is the case that minor (and occasionally major) decisions litter the field of casual sociality. In a conversation, a speaker may always get up and leave, and we typically are interested in this decision. A speaker may answer us in several different ways, and we are typically interested in which way he answers us. There is even the decision of whether he will answer us at all.

I fully agree with Natural Hazard that “affective manipulation”—the alteration of others’ mental states, and not behaviors—is often the most pragmatically useful level of abstraction, for an individual engaged in conversation. This is because, in many cases, we are unable to control the specific behaviors of a receiver (or may not want to) and are, rather, trying to ward off or make less probable an undesirable (to us) set of possible behaviors from our interlocutor, and make more probable a desirable set of possible behaviors from him. For instance, when our interlocutor begins to become upset, we may recognize that most outcomes of his upset are against our interests, and try to steer the conversation back to a place where his mood improves. At this level, it would be insane to model all the potential behavioral outcomes in an enormous decision tree and choose between. Instead, we proxy for these behavioral alterations via mental state. But we should not confuse the surrogate for the thing surrogated.

I do not think the mystery of casual sociality is solved in this post. I cannot analytically reduce a given conversation of casual sociality between friends to manipulation, the way I can the quartermaster’s orders, or the poker player’s display of post-flop arrogance. But given the arguments set forth—given that the evolutionary origins of language clearly identity it as manipulative; given that group decision-making and individual conflict scenarios are permeated with manipulative communication; given that we can only deal with the surrogates of mental states and are “trained” both evolutionarily and individually on the behavioral products of mental states; given that we already accept the idea of information as “difference which makes a difference”—I think the burden of proof rests clearly on those defending the idea that communication in this realm is not behaviorally manipulative, to make that case.

In the meantime, I think we ought to assume by default that, even though phenomenologically it does not always feel like we are performing behavioral manipulation, we in fact are, and in fact are optimizing around behavioral manipulation, and are in fact tailoring our interactive choices to behavioral alteration of our receiver, either locally or globally. Phenomenology simply is not enough to upturn the ACiM thesis: If we received honest reports from dancing bees that they were dancing “for” joy, and not to communicate the location of a potential colony nest to other bees, we would not suddenly toss out the claim that the bee’s dancing was a form of communication that could not be adequately understood without reference to its function in getting observing bees to fly to the signaled nest location.

* I believe that intentional unconscious communications occur, and that their intentionality (as well as, typically, the desires or goals they advance) can be gauged by their repeated, regular deployment in similar situations, and by the advantageous results of their deployment, but the thorny question of intent is mostly separate from the larger relationship between communication and manipulation.