‘Influence’ versus ‘coercion’

“If I disagree with the substantive theory assumed by your word choice, you can’t expect me to build that substantive theory into my language, because if I were to, I would be being forced to lie or contradict myself every time I use your term.”
– Sarah Fitz-Claridge


From the archives: The original post was posted on 1st January, 1995

A poster wrote:

“I agree with Sarah about the evils of coercion, but I’m afraid I’m in the camp she considers to have an unduly narrow definition of the term. In defense of my use of the word, I’ve consulted 3 dictionaries thus far, including Black’s Law Dictionary, and all agree that an essential part of ‘coercion’ is physical force or the threat thereof. As an alternative, I suggest she substitute ‘influence’ in place of ‘coercion’ to refer to all the things she wants to include in the category.”

Another correspondent said:

“‘Coercion’ (in plain English) is the use of force. You use it for not being the child’s obedient servant/slave. That is a ridiculous usage—whatever the merits of your teaching theory.”

I don’t think coercion in plain English is just the use of force. For instance, it is used for describing strikes: when people go on strikes, it is often said that they are trying to coerce the employer, or that the employer is trying to coerce them by changing the terms and conditions of their employment or whatever. When somebody at Ford steals some machine parts, and is given some punishment, such as suspension without pay or demotion or something, that is called a punishment, and the intention is (justified) coercion, even though there is not force or threat of force.

“You should write in English in order to be understood and to be honest (by avoiding calling things ‘coercive’ just because you disapprove of them). Your usage is confused and tendentious newspeak that obscures the real (entirely separate) theory—which you need to state in plain English to be better understood and to encourage criticism.”

This is not true. Language is not neutral. It can’t be. Language, like observation, is all theory-laden: everybody uses language in a way that partly expresses their theories. Language is a theory. When one changes one’s theory, one tends to change one’s language as well, because it becomes awkward to express a theory in the wrong language. One should, I accept, always be willing to express one’s ideas in the language of the person with whom one is conversing, but that does not mean that how one then expresses the ideas will sound natural.


I have a Marxist friend who insists on calling everything to the right of the Labour Party “fascism”. Okay, if he insists, I can call it “fascism” too; but in expressing my own views in other situations, with others who do not share the view that everything to the right of the Labour Party is fascist, it would be silly to do that. This is because, if nothing else, I am interested in fine detail of different types of so-called “fascism”, and it just makes more and more words in the sentence if I have to use the term in the way that my Marxist friend uses it. How am I to distinguish between Tory bigots and classical liberals, say, if I must call everything “fascism”?

So it is natural that I should use a term that expresses my theory, and you use a term that expresses your theory, but it is irrational of you to insist that I use your term just because you don’t like my usage, and it would be even more extreme if you were to refuse to offer an alternative, as some have.


I have explained in some depth what it is I am talking about. That piece was mainly about what I mean by “coercion”. So let me ask, what is this thing I am talking about in your language? The trouble with the terminology you suggest I use—“being the child’s obedient servant/slave”—is that it gives the same name to two different things. “Being the child’s obedient servant/slave” would cover a lot of things that are not what I am talking about. This point also applies to Tim’s idea that I should use the word “influence” instead of “coercion”.

This whole criticism does not make sense in view of the fact that the whole article was about the different ways in which people use this term. If I had only used one sense of this term in the article, I could not have written the article, because I would have had to say “For some people, “coercion” means so-and-so, and for other people, “being the child’s obedient servant/slave” means so-and-so.” That would not have made sense. But since you want me to use the latter term, perhaps I should.

(It is not easy to be both confused and tendentious at the same time. If one is to be properly tendentious, one can’t really be confused.)

You said:

“I merely wish you to use English instead of perverting language for propaganda purposes. You do not use force on a child unless you use force on a child. Not giving a child what it wants is not using force on a child.


I want use to use English instead of propagandaese.

All use of language is ‘for propaganda purposes’.”

BTW, for fun, I looked up “coercion” in the Complete(?) Oxford Dictionary. “Constraint; restraint; compulsion; the application of force to control the action of a voluntary agent.” “…a noble child, by his own natural disposition and not by coercion, may be induced to receive perfect instruction in these sciences”; … J.S. Mill: “the moral coercion of public opinion.” There were several uses of it that don’t involve force.

(You would no doubt say Mill was speaking metaphorically, and I accept that.)


When we are talking about educational theory, we are talking about the application of epistemology to psychology, and we need a vocabulary to talk about the different types of states a person might be in—pain, satisfaction, preferences—there is a great vocabulary about this. And we can also talk about interactions with the outside world, and we need to classify those as well. This is because when we are trying to apply epistemology to these things, we are dealing with subtle distinctions, and whilst it is certainly an important and useful distinction whether there is force or threat of force or not, that does not begin to cover the different types of structure that a relationship can have.

Insisting that there is only force or not-force is not useful, because it takes no account of the many different ways one could experience force or not-force. Perhaps we are arguing on different levels here. My argument is not about coercion as defined by the law, but about the effect on a person’s mental/psychological state of certain ways of interacting with him. To insist on the use of “coercion” to describe only physical force or the threat of force is to deny that certain types of interaction can cause a person to be pressured into doing or not doing something whilst they have real wishes in the opposite direction. I repeat, this is not an argument on a legal level.

I want to try to illustrate my proposition that there are things relevant to human affairs other than whether of not an action is legal or not, by an example that may be more natural to you than looking at coercion of children. You might like to substitute examples of parental treatment of children.


Suppose a wife is beaten up every day but is free to leave. In law and ethics, she is not being coerced; she is consenting. [Note added 2022: That is no longer the case in English law, BTW. Now the law even recognises “coercive control” which might not involve any physical violence.] It does not matter what you call it, but if you are studying psychology, you have got to understand that that situation is different from a happy marriage. You have got to use different terms. Just because the law makes no distinction, that doesn’t mean that there is no difference in the two situations. Ordinary language does use the word “coercion” to describe that first situation. Some libertarians would (to be consistent) say that this situation is exactly the same as a boxing match, and that there is no coercion going on here, but I want to draw a distinction between the way the boxer experiences being punched in the match, and the way this woman, who is as free to leave her marriage as the boxer is to leave the ring, experiences being punched. The boxer is not in any psychological conflict. The wife most certainly is (assuming this is not part of some completely consensual sado-masochistic activity).


Now at risk of repeating myself ad nauseam, I am not saying that I do not understand the legal or libertarian use of the term, or that that usage has no value; I am merely saying that we are talking on different levels here.

In this marriage, there is an unpleasantness in this psychological situation and a complicity, to which we wish to give a name. (Just like there is an unpleasantness in presenting a child with two options when the possibilities are endless, and calling that “freedom of choice”, or deciding what you think the “natural consequences” of a given action are, and “allowing” the child to experience them, when actually, that is all a lie.)

To get back to this unpleasantness in this psychological situation, let’s call it (as you might have me call it) “not giving one’s wife everything she wants”. (not “being the child’s obedient servant/slave”) What’s in a name? Okay, we’ll call it that. That means that some husbands are very culpable for not giving their wives everything they want. This husband is beating his wife, remember.

Then immediately we run into another problem: there are some situations in which the husband does not give his wife everything she wants, which we would not consider culpable. Others which we would. If we are going to discuss the rights and wrongs of marriage situations in which the parties all behave legally, if we insist upon using legal language, we are going to describe everything they do in the same words. We may as well just say everything they do is called “behaving legally”. The husband gets up and behaves legally, He goes to work and behaves legally and later comes home and behaves legally. They go to bed and he behaves legally. (Under libertarian law.)

I want to say that the husband is doing right, doing wrong, he is behaving badly or not, he is a cad, he is not a cad, he is being reasonable/unreasonable, she is a termagant, she isn’t a termagant, and so on. We don’t want to discuss merely the legality of their actions, for example. We want to give meaning to all those statements—to apply some morality. But let’s stick with using “your” terminology: some cases in which the husband does not give the wife what she wants are reprehensible morally and others aren’t. If the wife were to come to you and complain about her husband’s behaviour, in some cases, you’d say, “God, that’s awful. You should leave him straight away,” whilst in other cases, you wouldn’t. The point is, there are many behaviours which are morally wrong which are legal, and there are behaviours which are morally right which are not legal.


This example of morality is just that—an example of something one might want to talk about when one is talking about human affairs—it is not the only example. One could talk about practicalities instead. The point is, there are things we shall want to talk about that do not depend upon whether the thing is legal or not. This proposition that there are things relevant to human affairs other than whether an action is legal or not is something some libertarians seem to find hard to accept. Ethical subjectivists may quibble about my use of terms like “morally wrong” but even ethical subjectivists do object to things, for their own reasons, so the point I am making still makes sense even within that ethical framework. Whether your objections to a given thing are subjective or objective, you do object to certain things, and you use as your criteria something other than whether these things are legal or not. What we are talking about is simply whether there is something to talk about.

For instance, in the language as we are currently using it, there are many different cases of the woman “not being given what she wants”. Within the category in which the man is perfectly within his rights not to give her those things, there are still many possible cases. For instance, some of them are happy marriages and some aren’t. We might want to analyse what makes a marriage unhappy. That has nothing to do with ethics/law. It seems obvious to me that when one is asking things like “what is conducive to happiness?”, the answer will depend upon how people perceive situations.


A crucial thing is what sort of external events cause, say, hang-ups, or unhappiness? It is the subjective experience of the action that counts, not whether the action is justified or unjustified legally or whatever. The only thing that counts is the person’s experience of it. For instance, a person might become very very unhappy if they think they have been cheated in a particular situation, when they have not in fact been cheated, according to libertarian ethics. A Marxist friend of mine just resents his employer on principle. He feels that to work is to be exploited, and he genuinely suffers as a result. If he were to have certain types of interaction with his employer, he might become outraged or suicidal or whatever—but what reaction he might have would be entirely dependent upon what he thinks the situation is, not on what it actually is.

Typical categories of discussing human affairs will refer a lot to people’s internal states and perceptions. For instance, if we are defining the term “aggrieved”, there is a legal definition of that which means that the person has a legitimate grievance, but there is also a psychological notion of “aggrieved” which means that the person believes that he has been hard-done-by. Obviously, the psychological effect of being “aggrieved” depends entirely upon that second definition and not at all on the first definition.

Suppose we talking about what makes people become terrorists—one of the very common conditions is that a person is aggrieved by society, by the government, etc. A libertarian might say, “Just a minute. That person was not aggrieved at all. In fact, he was a net beneficiary of the government. He has no grievance at all.” But that is neither here nor there. It is not the case that just because he is not aggrieved according to libertarian ethics, he won’t become a terrorist. It is his subjective perception of the situation that is important, not the legal details. I don’t know if I am explaining myself very well here. All I am trying to argue is that insisting upon narrow legal word usage is unhelpful, because it masks a world of interesting differences that are worth thinking about.

To come closer to the word “coercion”, let’s look at the notion of “voluntary”. The psychological use of the word is quite different from how it is used in law. If somebody comes to you and takes a job, you offer them something and they accept, then from the legal point of view they are working for you voluntarily. Of course in such a situation, my Marxist friend might consider himself a slave, and might think that you are exploiting him. In fact you are not exploiting him, but his inner life is exactly as if you were.

There is an important difference in the psychology of such a person from that of someone else who does not feel like that. Ordinary usage here agrees with what I would say and what psychology would say, which is that my Marxist friend, who believes that the employer-employee relationship is corrupt—and who, in his scheme of things, sees the employer as a thief—is not working voluntarily. We can define him as working voluntarily (just as we could define a “choice” between tidying away toys or having them put in the attic, as freedom of choice), but then we would be casting him into the same category as a libertarian who thinks that the employer is doing him a favour. How are we to talk about those people? Can you really not see any distinction between the two cases? I think ordinary usage does use these words subjectively, when necessary—especially since ordinary usage is not libertarian.

Suppose you have a friend who is a psychologically unstable Marxist, and for some reason you are keen not to put him in certain psychological states that you consider dangerous. Suppose that you think that if he has a grievance against somebody, he might murder them. (Just suppose that this were the case. I am not saying it is at all likely or wise—the details of this fantasy are irrelevant—it is just by way of illustrating a point about language, etc.) That might lead you to think twice about whether to employ him or not, otherwise he might kill you. If for whatever reason, you wanted to avoid this state of mind in him, you’d also have to make sure that nobody employed him. You would want to make sure that everything he did was of the form that did not make him think he had a grievance that someone was doing something to him illegitimately against his will. Or to put that more concisely, you’d have to make sure everything he did was voluntary.

With such a person, this would not in general be the right thing to do—the thing to do might be to let him stew in his own juice and give up on him—but if you did want to avoid this state of mind in him, that is what you’d have to do—you’d have to consider what is voluntary in the subjective sense of the word, not the legal sense. If you did not want to use that word “voluntary” to describe what you were trying to do with this person, you would still have to do that thing, and if it was an important part of your life to do that thing, and if you did that thing, you’d soon find yourself inventing a word for it.

As I said, the point is nothing to do with the wisdom or otherwise of pandering to the irrational ideas of psychologically-unstable Marxists: it is that there are things we might want to talk about or bear in mind when acting, that do not depend upon legal word usage. 


The theory that coercion means violence is a mistake. What about blackmail? Suppose someone of a certain uptight social class is gay but married, and does not want it to be known to anyone, and that if he were found out, he might not lose money over it, but he would be looked down upon by those who currently look up to him. So this person is now in a psychological conflict: he does not want to give in to blackmail, but the thought of being exposed is traumatic psychologically. If he gives in, he will be acting upon a theory (to hand over the money) whilst another theory (that he does not want to hand over any money to this lowlife) remains outstanding, unrefuted. If he does not give in, he has the trauma of exposure. Either way, he suffers. Either way, as a result of this blackmailer’s action, he has an unpleasant internal conflict he would not otherwise have had. You could define these choices the blackmailer is giving him as “free”, and the consequences of either choice as “natural”, but we all know what is really going on there, don’t we? Yes, some people would call blackmail of such a person legitimate; but others would call it coercion. It is just blind to say that ordinary language does not call that coercion.

It is for similar reasons to those given in this blackmail case that I say these “natural consequences” parents are often so keen on, or fake “free choice” they give their children, are in fact coercive: they put the victim—err—child, I mean—in an impossible position; they raise intractable internal conflicts; they force the child to act on one theory while another (or others) remains unrefuted—while another idea remains still attractive to the child, that is.

You are insisting upon defining coercion in terms of external events, rather than in terms of the internal state, so “not doing what the child says” may or may not be coercion in my sense. That is why calling it “not being the child’s obedient servant or slave” is very, very unrepresentative of what I mean. I’d have to say something like “I am the child’s obedient servant where doing otherwise would distress him, but I should hope to not get into that situation in the first place, by reaching mutual consent.”


Another thing that is wrong with that definition is that it is asymmetrical. It is true that non-coercion involves thinking about what the child wants, but reaching a real solution is also doing what everyone wants. Whatever the rights and wrongs of it, the practice of giving in to the child in cases in which the parent-child knowledge-building entity does not have the requisite creativity to find a solution—that case does not look like a master-servant or slave relationship. In a master-slave relationship, the wishes are a priori fixed, and the whole relationship consists of the execution of one person’s wishes by the other one. Can you not see the difference between any given master-slave interaction, and failure in a given case to reach consent in a relationship whose default is consent-based decision-making?

You might say that if one starts doing this—giving in to the child when one can’t find a real solution—a master-slave relationship is what it will deteriorate into. But that is a substantive theory. If I disagree with that substantive theory—as I do—you can’t expect me to build that substantive theory into my language, because then every sentence would have to be something like “I am going to do what the child says, except that the child won’t say that, he’ll say something else.” It is a bit like saying that the scientific community operates by everybody being a slave to the successful theory. It is a twisted view of things.

See also:

Sarah Fitz-Claridge, 1995, ‘“Influence” versus “coercion”’, https://takingchildrenseriously.com/influence-versus-coercion/

Leave a comment