Educational theory: science or philosophy?

“How to treat children is a moral issue, not science. Educational theory is epistemology applied to psychology: its subject matter is the contents of minds, and you can’t control for all the variables—the ideas—in people’s minds. The number of ideas that a single mind could hold is far greater than the number of people on Earth, so it doesn’t matter how large the sample—it won’t be enough.”
– Sarah Fitz-Claridge

From the archives: Originally posted on 2nd March, 1996

Note (2025): Parenting experts sometimes scare parents by making it sound as though Science Says coercion is necessary for children. Such fallacious arguments from the authority of science are often couched in language like “evidence-based, data-driven parenting strategies backed by science.” (As though the moral issue of how to treat children could be determined by a scientific study!) Such assertions today are like the assertions of experts a hundred years ago, who declared in many books I have read from that time, that Science has shown that certain peoples are inferior and must be (forcibly) sterilised, for their own good of course. Eugenics was as popular at that time as all this coercing children in the name of “science-based parenting” is now. If you are a parent worried about going against the experts and their alleged evidence-based this, and science-backed that, you might find the following posts distinguishing between science and philosophical/moral/psychological theories reassuring:

Educational theory: science or philosophy? Post 1

One of the main obstacles to understanding educational theory and in particular noncoercive education [and taking children seriously in general], is the conventional wisdom that philosophical theories—ideas not testable by empirical means—are worthless by their very nature, and the related misconception that educational theory [and parenting] is scientific and amenable to testing.

First, what distinguishes science from non-science?

The aim of science is truth: true explanations of reality. But astrology also purports to be true, so what is the difference between this and science? What characterises theories of science as opposed to non-scientific theories is that scientific theories are experimentally testable. What this means is that, to be scientific, a theory must make predictions which can be tested and in principle refuted. So a scientific theory has to forbid something. (There is no point doing an experiment if that test could not possibly show the theory to be false!) This is why astrology is not scientific. But this is also why theories about human psychology are also not scientific (despite most psychologists’ views). And why the theory of evolution is not scientific in itself (although specific theories within that philosophical framework are scientific).

This idea of falsifiability as the criterion of demarcation between science and non-science has been taken by some to be a criterion of demarcation between science and nonsense. Indeed, in the case of pseudoscientific mumbo jumbo like astrology, I can see why. Some people say that science is everything and everything else is nothing. They are effectively saying that if it can’t be measured, it doesn’t exist. But this is a mistake. Philosophical ideas such as the theory of evolution and the theory of knowledge are not empirically refutable and yet are certainly not “nothing”. That, and not science, is the category that educational theory is in.

But doesn’t applying this “non-science” label denigrate evolution and educational theory? Surely the theory of evolution can’t be on the same level as pseudoscience like astrology. Of course not! The framework of evolution, and indeed many other philosophical frameworks such as realism, for example, have been more fruitful than any scientific theory. Indeed all scientific theories rely upon these philosophical frameworks. Non-scientific theories can be just as much a part of the search for truth—for true explanations of reality, for example—as scientific theories. They may not be refutable empirically, but they are still true or false, or better or worse.

This anti-non-science attitude seems to lead people to try to squeeze everything into “science”, whether or not the theory is empirically testable, and unfortunately this grave error bears directly on ideas in the area of educational theory and parenting. Empirical testing is a powerful tool in the realm of science, but of no use whatever in areas whose theories are not empirically refutable in principle. But if non-scientific theories are not testable empirically, how can they ever be shown to be false? Wouldn’t it be better to make them experimentally falsifiable? No. You can’t make an inherently untestable theory testable just by an act of will. In these non-scientific areas we need other forms of criticism—argument, not experiments. Non-scientific theories stand or fall on whether they survive argument and whether their explanations are consistent with other ideas, better than the explanations of theories they replace, more fruitful as frameworks for testable theories, and so on.

One major reason people make the mistake of trying to squeeze philosophical / psychological theories into science is that they hold the view that there is objective truth in science but not in other realms. Since they rightly want to believe that, for instance, the theory of evolution is objectively true, or that there are better and worse ways of bringing up children, or that there is some value in some ideas in the field of psychology, they think that they also have to take the view that these ideas are scientific.

If instead one takes the view that there is objective truth in areas other than science then one can see that experimental testing is just one of a number of forms of criticism, argument being another, and that experiments are only useful insofar as they can in principle show one or more contending theories to be false. This view also allows one to see that scientific theories embody (or at any rate are in a context of) philosophical (and thus non-falsifiable) ideas (realism, for example). Indeed, scientific theories embody even ethical standards. This is only a problem if you take the view that there is no objective truth in ethics. Perhaps I should stress that believing that there is objective truth is not the same as believing that we know what the truth is. Just as, in science, our ideas are improving as old theories are overturned by better ones, ideas can improve in other areas too.

Educational theory: science or philosophy? Post 2: The framework theory of evolution is not scientific. So what?

I have said (Post 1, this series [above]) that some of the main impediments to understanding the non-coercion idea are misconceptions about the similarities and differences between scientific theories and philosophical theories. I shall elucidate my point in part by criticising the arguments of two posters in a debate on the Home-ed List about science, philosophy and evolutionary theory.

Joe wrote:

“Science is concerned with the process of building models which describe nature to the best extent allowed by current experimental accuracy.”

It is true that the purpose of science is to build models which describe nature, but to say “to the best extent allowed by current experimental accuracy” suggests that by “describing” nature you mean predicting. Describing nature is largely not predicting: it is about explaining nature. Predictions follow logically from explanations.

Moreover, many of our theories are overwhelmingly better than experimental accuracy. Scientific theories are not induced from observations and experiments; they are bold attempts to explain reality, and they are not the product of any mechanical process but of human creativity. The theory always comes before the testing, and indeed sometimes theories cannot be tested immediately they are created, because of the limits of technology, for example.

Joe has suggested that the theory that coercion diminishes creativity could be tested experimentally in principle. The reason this notion is mistaken is given in the post [see below] “Educational theory: science or philosophy? Post 3—Joe and Heather revisited.” See also other posts in this series.

Rick wrote:

“Evolution is usually considered a “belief” precisely because experimentation—and observation—are both impossible with regard to finding out the actual origins of the various species on earth. Since observation and experimentation are both so essential to science, it is hardly appropriate to consider anything “scientific” where these are both impossible to be applied.”

Evolution is a philosophical theory which really amounts to saying that the world is not supernatural. If the world is not supernatural, then it follows essentially by logic, that complexity must have arisen through evolution. (Complexity is evidence that some computation has happened. Computation is a physical process which transforms some input information into some more useful output information—it gets an answer that satisfies a criterion where that criterion is hard to satisfy in that not all answers would satisfy it.)

To me, “supernatural” means that there is complexity that is not physically explicable. So the theory of evolution says that all the complexity we see around us is physically explicable in principle. Clearly, that can’t be tested against a theory that says the complexity around us is of supernatural origin, because the supernatural origin could always be such as to mimic any evidence that we might present; therefore, evidence can’t test between those two theories. So in that sense, the theory of evolution is a philosophical theory and is empirically irrefutable.

However much I might like this not to be so, solipsism (the theory that the self is all that exists or all that can be known) too is empirically irrefutable because the solipsist can just say that any evidence presented is all in his mind or whatever. Solipsism must be rejected on philosophical grounds. No amount of empirical evidence is enough to refute it. This is only a problem if you think that empirical evidence has authority in a way that philosophical arguments do not. There are some very powerful philosophical arguments against solipsism.

When Karl Popper said that evolution is a metaphysical theory, everyone jumped on him, and he later tried to put it in different language, describing evolution as a philosophical “framework” or something. But that does not mean that it is optional! In other words, it is not, I conjecture, a philosophical framework such that someone else might have a better philosophical framework. A philosophical framework is a framework within which we construct specific theories. Those specific theories are testable, although very hard to test. For instance, one might have a theory that the evolutionary advantage of sexual reproduction is that it mixes genes, and then one might test that by testing first of all whether genes are mixed when species undergo the earliest forms of sexual reproduction such as gene transfers.

But it is fallacious to say (as do some who believe that the universe was created in 4004 BC) that evolution is “just a theory”. It is an equivocation designed to disparage the status of evolutionary theory by implying that it is an irrational belief rather than a true theory having deep explanatory power that the 4004 BC theory lacks. As Joe has said wisely, all our best ideas are theories. So what?

Rick wrote:

“Evolution is usually considered a “belief” precisely because experimentation—and observation—are both impossible with regard to finding out the actual origins of the various species on earth.”

Depending upon what you mean by that, it is either a truism without much importance, or it is false. The origins of particular species on Earth are testable scientific theories, for instance that humans are descended from apes is testable by matching our DNA with that of apes. It has frequently happened in evolutionary biology that it was thought that one species was descended from another, but when we have examined their DNA, we have found that they were too far apart for the one to have descended from the other, and that actually it was descended from another species that looks completely different. In the case of human beings and apes, we could hardly be closer related, but we are different enough to know that we can’t actually be descended from present day ape species: we are their cousins rather than their descendants. So in fact in its fine detail the theory that humans are descended from apes is not only testable, but has been tested and refuted.

Joe replied to Rick:

“That’s clearly nonsense. We observe massive numbers of fossil remnants of animals and plants that once were. We observe a striking pattern of radiation amongst these fossil remnants. it is very difficult to have even the most basic knowledge of geology and cling to superstitious beliefs about the origin of life on earth.”

True, but none of that makes the framework theory of evolution testable. All these things you are pointing out are confirmations of the theory, and thus do not make it a scientific theory. It is a fallacy to say that because we observe massive numbers of fossil remnants the theory of evolution must be true; because the rival theory—namely that God created those fossil remnants—also predicts that they will be there. The rival theory must be rejected on philosophical grounds, not because it has been tested. No number of fossils will refute it, only philosophical reasoning.

It is not true that “…experimentation—and observation—are both impossible with regard to finding out the actual origins of the various species on earth,”as Rick suggests. Experimentation can address the actual origins, given the philosophical framework (e.g., the humans and apes stuff).

Joe again:

“We can experiment and produce new species in the laboratory.”

I know we can produce new variants of species and new strains of breeds but I am not aware of actual new species produced in the lab (but if we can’t now we soon will be able to).

Joe again:

“We can observe new species evolving in nature.”

I am not aware of this but I’d be interested to hear more.

Joe again:

“We can observe the striking patterns of anatomical variation amongst extant species.”

Yes, for that matter we can observe evolution amongst extant species, for instance, the famous butterflies that turned black because of industrialisation. (There is a species of butterfly which used to be white but slowly turned black over a couple of hundred years, probably because the trees on which it bred turned black because of industrial pollution. I don’t know but I should imagine that they are slowly turning white again. )

Rick wrote:

“Since observation and experimentation are both so essential to science, it is hardly appropriate to consider anything “scientific” where these are both impossible to be applied.”

Isn’t this the fallacy of division?

Joe replied to Rick again:

“This is silly. You’re creating a very nutty definition of science, which would rule out all of cosmology, the study of the history of the universe. We can’t recreate universes in the lab. The evolution of the universe as a whole is a one time thing. All we can do is observe what is here now, and infer how it got to be this way as accurately as possible.”

Cosmology is scientific. Different cosmological models make different predictions about what the world would be like today. Cosmology is not the same sort of theory as the evolutionary framework. Rick’s statement doesn’t rule out cosmology any more than it rules out any other part of science. (You are both wrong!) All science is in the context of philosophical frameworks such as realism, evolution, physicalism, or whatever.

If Rick were to think that lightning is God’s punishment for being wicked, and you were to point out that actually lightning occurs when the voltage between the cloud and the earth exceeds a certain level, he would simply say that God decreed it that way, in precisely such a way that it would punish the right people at the right time. So seeing lightning and seeing that it obeys the theory does not tell you that the rival—God—explanation is false. In other words, you have to already think that the philosophical framework is true in order to believe the explanation. (That is what I meant by saying that scientific theories are in a philosophical context.) But the philosophical framework is not less convincing for being “merely philosophical”. On the contrary, it is not an entrenched irrational belief; it is still open to criticism.

Furthermore, these kinds of frameworks are among the most severely tested in the critical sense; they have no real rivals other than religious ideas. Some of these frameworks are philosophical beliefs which are better tested, better corroborated, in the sense of having survived criticism, than any scientific theory. So they shouldn’t be downgraded because they are not scientific—they are actually better than scientific theories. All scientific theories rely on them.

Joe again:

“Similarly, from the pitted surface of the moon, we can infer the characteristics of the meteorites which have bombarded it in the past. According to you, such study is not scientific, because we can’t recreate the moon and its history.”

That isn’t so. We can test the theory that the meteorites had a certain property in the past other ways than by creating another moon. For instance, if that property was the density of meteorites over the past few billions of years, then we might ask how would that density have affected other planets with say atmospheres, or different sizes or different gravities? Then one could look at those planets to see whether they have the pattern that they would have if that had been the density of meteorites. If they haven’t, the theory from which that prediction followed is refuted.

Joe had written:

“That’s clearly nonsense. We observe massive numbers of fossil remnants of animals and plants that once were.”

Rick replied:

“There are also “massive amounts” of relics and writings. That does not make religion scientific. Experiments can even be, and often are, performed on relics to determine their origin. That still doesn’t make it scientific, just as looking at fossils and experimenting with them doesn’t make evolution scientific, because its foundation is unobservable and beyond the realm of experimentation.

Looking at relics is scientific (or can be).

Joe had written:

“According to you, such study is not scientific, because we can’t recreate the moon and it’s history.”

Rick replied:

“That’s right, because though hypotheses can be formed, they can never be tested to determine whether or not they describe what actually did happen.”

It seems to me that when you all say “test” you mean “confirm”.

What I have tried to show here is that one of our best theories (the framework theory of evolution) is not scientific, but that it is none the worse for that, and that scientific theories all rely upon a philosophical framework.

Educational theory: science or philosophy? Post 3: Joe and Heather revisited

[You may want to read this post which I quote from in the following.]

This is one of several messages I am posting on the nature of educational theory, because it seems that there is much misunderstanding about what constitutes scientific theory, and I want to show that theories about human psychology, etc., are not scientific and can never be scientific theories. I want to explain why (interesting) theories about human learning cannot be tested empirically, and (see other posts in this series) show that philosophical theories are nevertheless vital. These posts, in other words, are an answer to suggestions that such-and-such a “counter-example” refutes the theory that coercion diminishes creativity, etc. In case it is not obvious, I should stress that I am not necessarily interested in the truth of the particular theories discussed here—I am just discussing whether they are scientific (empirically testable) or not, in an attempt to explain why I brush off these “counter-examples” people keep throwing at me.

BTW, I know that what I am saying is not obvious, judging from the number of “experts” who say exactly the sorts of things I am criticising here. So, Joe, I am not knocking you, I am just using your post(s) because I know you love a good argument and you happen to have said all the “right” things.

[x] I added these numbers in square brackets to the text of Joe and Heather to highlight the theories.

Quoting from this post:

Joe had written:

“Heather was arguing that because her kids

[4] learned to walk and talk in inverse proportion to the amount of teaching she did,

that meant that

[3] teaching walking and talking was generally ineffective.”

He suggested:

“I offered evidence to weaken that conclusion.”

I had replied:

“Hmm, well let’s look at what Heather really said in that earlier message:

[Quoting Heather:] “When my older daughter was a baby, I thought

[1] I had to do something in order for her to learn to walk and talk.

I did baby exercises with her, deliberately talked to her all the time and encouraged her to copy me and fussed a lot. She was an average walker and an early talker. When her sister cam along, I didn’t have the time to worry about teaching her to walk and talk. She walked at 9.5 months and was an earlier and more effective talker than her sister.”

Joe had replied to Heather:

[2] “I did nothing with my first 2 kids, and they were late walkers and talkers. I did much more with the baby, and he was a very early walker and talker.”

I had commented:

“Between them, Heather and Joe have touched upon the first four of the following five theories:

[1] Teaching is essential for learning to walk and talk
[2] Teaching assists learning to walk and talk
[3] Teaching is irrelevant to learning to walk and talk
[4] Teaching is harmful to learning to walk and talk
[5] Teaching prevents learning to walk and talk

Specifically, Heather began by stating that she used to hold theory [1], but that her subsequent experience refuted that theory. It does refute it (but we must ask, did Heather really hold it, or was she exaggerating?).

Then Joe accused Heather of asserting theory [4], and theory [3]. (He can’t have it both ways, though.) Perhaps she was asserting one of them, implicitly. However, even if Joe is right that Heather is asserting [3] or [4], his reply is irrelevant because his observations are perfectly consistent with theory [3] or [4] being true.”

Joe replies:

“Hmm? The fact that my children learned to walk and talk in proportion to being taught is perfectly consistent with 4? That’s bizarre. Clearly, this kind of evidence strengthens 1 and 2, and weakens all the other theories.”

No it doesn’t! This is false. It is consistent with 4, as I said, because you do not know what would have happened had you not taught the children. Theory 4 was that teaching is harmful to learning to walk and talk, not that it prevents learning to walk and talk. The children might have learnt to walk and talk more quickly had you not taught them (see Post 2 [above]).

Now I can see that at first glance, this sounds untrue. I can see why someone would think I am talking rubbish, so let me explain further. One theory everyone is familiar with is that smoking is harmful to the lungs. That theory does not imply that every smoker will die young: some smokers will live to age 99, and that is perfectly consistent with the theory. We do not throw out the theory that smoking is harmful whenever anyone points out that their Uncle Jim is still alive at 99 and still smoking 40 a day. That is because logically the theory that smoking is harmful also predicts that some smokers will never get lung cancer or any smoking-related disease, and will live to a ripe old age. This is not saying “smoking will cause the person to live to a ripe old age”; it is saying that some smokers will be lucky, because of good genes, or other unidentifiable factors outweighing the effects of smoking. There are many possible factors, and the theory that smoking is harmful recognises this.

Similarly, the theory that teaching is harmful to learning to walk and talk predicts that some children, despite being taught, will learn to walk and talk faster than the average. If it predicts this, it can’t be refuted by it! The teaching might be offset by some beneficial factors which in that case outweighed the harm done by the teaching. There are many factors! Suppose that teaching actually delays the onset of walking and talking by three months, but that being out in the sun advances the onset of walking and talking by four months. Then you will be able to point to a child who learned to walk and talk early despite being taught. That is the difference between the harmful and “prevents” theories.

So, in answer to:

[Quoting Joe again:] “Hmm? The fact that my children learned to walk and talk in proportion to being taught is perfectly consistent with 4? That’s bizarre. Clearly, this kind of evidence strengthens 1 and 2, and weakens all the other theories.”

No. It is not evidence; and it doesn’t. Evidence has to be something which distinguishes between two of the contending theories. This doesn’t.

I had said:

“Nor does it ‘weaken’ Heather’s refutation of theory [1]. Nor does it strengthen theory [2], the theory that Joe is implicitly endorsing, because evidence confirming a theory does not strengthen it.”

Joe replies:

“Rather, I suppose, like the cosmic background radiation doesn’t strengthen the Big Bang theory. Perhaps you might want to contact the Nobel committee to withdraw the physics Prize from all the people who’ve won it for evidence confirming various theories. You have a very striking and unconventional view of evidence indeed. You might want to relay this new insight to the courts as well.”

An example of evidence which “confirms” the Big Bang theory is that stars which are further away seem redder. But that is not evidence in favour of the theory, because unfortunately it was also predicted by the rival theory (at that time, the Steady State theory). And since it was also predicted by the rival theory, it was not evidence in any useful sense. Cosmic background radiation was evidence that refuted the Steady State Theory, not confirmed the Big Bang theory. One could propose an ad hoc hypothesis that there is a new law of physics, in addition to all the other laws, that predicts background radiation with the Steady State theory—in order to rescue the Steady State theory—and that is indeed what someone did say (falsely). If the Steady State theory had also predicted that there would be microwave background radiation, then that would not have been evidence. Similarly, somebody writing in to a medical journal and saying that his grandfather is 99 and has been a heavy smoker all his life doesn’t weaken the theory that smoking is harmful (see post 2).

You don’t get the Nobel prize for evidence confirming theories, you get it for making a new discovery. But the epistemological argument as to whether it is confirming the theory or refuting the opponent is not relevant here: what we are talking about here (the teaching theories 2, 3 and 4) is a case where the theory and its opponent both predict the same thing, and that is not evidence for or against anything. The question here is not when you have genuine evidence, is it genuine evidence because it confirms something or because it refutes something? That is a different question. This isn’t genuine evidence.

Joe, you have simply made a mistake: the things you are so sarcastic about are in fact true, as you will realise if you think about it more carefully. You asked sarcastically: “The fact that my children learned to walk and talk in proportion to being taught is perfectly consistent with the theory that teaching is harmful to learning to walk and talk?” The answer is yes.

I had said:

“Theories [1] and [5], the uninteresting ones that no one really believes, are the only refutable ones, and indeed both are refuted by the evidence of both Heather and Joe. The other three are not refutable, since there is no possible outcome that they predict won’t happen. (Since a controlled experiment is impossible here, we cannot tell what any of the children would have done if they had been treated differently, because we cannot hold constant any number of other factors that could be involved.)”

Joe replied:

“Well, we could, but it would be prohibitively expensive and there would be the usual plethora of troublesome ethical questions that surround experimentation on humans. There’s no reason to believe that this kind of study can never be achieved at some time in the future, though.”

I’m afraid there is. The way to hold all the variables constant would be to use similar techniques to those psychologists employ when the subjects are rats. One would have to use genetically cloned children—so that they are genetically identical; in addition they would have to grow up in identical environments, i.e., identical boxes in the lab, as with experiments on rats. One could not allow any interaction, because immediately the environments would become non-identical. Remember that the variables include the contents of people’s minds.

Suppose you did do all that, and you got some results: the results of that would not tell you anything, because they would only tell you how people brought up in those circumstances behave. They would not tell you the effect of, say, living in a normal society; they would not tell you how that might change the results at all. So there is a fundamental reason why one can never do such an experiment on humans. Because if you controlled all the variables, you’d be experimenting on them in a situation that is uninteresting. And if you did not do it that way, the variables would not be controlled. (See posts 4 and 5 in this series [below].)

I had said:

“I have set out this little analysis because this invalid form of argument-from-experience often appears in debates about education and child-raising.”

Joe replied:

“All these theories could either be refuted or receive strong evidentiary support from the systematic analysis of very large collections of human experience. There are billions of people on this planet, all of whom were once children. As we build up the capacity to record and process information, there is a wealth of data out there for us to mine. I’m not as pessimistic as you are about our ability to gain insight via experience.”

This has got nothing to do with optimism or pessimism. If somebody asserts that something is false or true, it is not a question of optimism or pessimism, it is a matter of logic. Ask yourself why the proposal is that “very large” samples should be used. Could it be that it is hoped that with “very large” samples, the variables will cancel out? The problem is that there are far more variables than people on Earth, because the variables include the contents of people’s minds.

I know that what I am saying here is not recognised by many people, indeed almost no one in the field of human psychology itself understands this point, and they all continue to think of what they are doing as science… Anyway, can you see what I am saying?

Joe said:

“You appear to live in a twilight zone where argument and theorizing are the only way to gain insight into…you guessed it…future experience. I have a more balanced point of view, where argument and theorizing must be informed by past experience.”

Past experience which isn’t relevant can’t be used to decide between two theories! No amount of optimism is going to change the nature of theories about psychology, whatever psychologists and educationalists like to think.

I had said:

“I would also like to stress perhaps a more important point about this exchange, namely that both parties make a completely unsubstantiated assumption that learning to talk and walk early is a good thing. This is a fundamentally mechanical criterion of educational success.”

Joe said:

“Well, my children were frustrated at not being able to move around like other people, and not being able to communicate their needs clearly. They sure seemed eager to increase their communication and mobility. But perhaps all children are not like that, and so early walking and talking might actually make them unhappy that people understand them better, and that they can now run around.”

This is shifting the argument. If your children wanted to learn to walk and talk then given the opportunity they would learn anyway. The issue was about teaching (which implies externally imposed instruction or a parental agenda if you prefer). If all this arose out of the child’s problem situation, fine, but when parents advocate teaching, that is not often what they mean.

Educational theory: science or philosophy? Post 4: Doris’s non-coercion theories

Doris writes:

“Sarah lists a continuum of theories:

‘[1] Teaching is essential for learning to walk and talk
[2] Teaching assists learning to walk and talk
[3] Teaching is irrelevant to learning to walk and talk
[4] Teaching is harmful to learning to walk and talk
[5] Teaching prevents learning to walk and talk’

What about this continuum of theories?
[1] Non-Coercion is essential for the acquisition of knowledge
[2] Non-Coercion assists the acquisition of knowledge
[3] Non-Coercion is irrelevant to the acquisition of knowledge
[4] Non-Coercion is harmful to the acquisition of knowledge
[5] Non-Coercion prevents the acquisition of knowledge”

I had said:

“Theories [1] and [5], the uninteresting ones that no one really believes, are the only refutable ones, and indeed both are refuted by the evidence of both Heather and Joe. The other three are not refutable, since there is no possible outcome that they predict won’t happen. (Since a control experiment is impossible here, we cannot tell what any of the children would have done if they had been treated differently, because we cannot hold constant any number of other factors that could be involved.)”

Doris replied:

“I agree. [1] and [5] are uninteresting. No one really believes them. They’re refutable. The rest I can live with. ;)”

Very amusing, Doris, but this is false, and there is a hint of equivocation here I think. If by “non-coercion” you mean specifically that the parent advocates non-coercion, then okay, I agree: not all children of parents who advocate either coercion or non-coercion appear to be completely uncreative. But what the parent thinks is not the issue: what matters is the coercion in the child’s mind. That is not a measurable thing, and nor is the growth of knowledge measurable. What theories the parent actually holds is also not measurable, if it comes to that—not even by the parent himself—because we all have inexplicit theories to which we do not have direct access, by definition. BTW, I disagree with the implication (suggested by the word “acquisition”) that learning is a passive process.

The child of a parent who does not hold the theory that coercion diminishes creativity may well nevertheless find avenues of creativity not cut off by parental coercion and thus learn—that is the norm indeed, to an extent, so coercive parents may or may not be relatively lucky in that respect. Conversely, it is unfortunate that, since the growth of knowledge is an active process on the part of the learner, some parents are unlucky: they may not be exceptionally severe and tyrannical and yet have a child who aims straight for the areas in which they are coercive and thus becomes autistic or something.

Educational theory is epistemology applied to psychology: its subject matter is the contents of minds, and one can’t control for all the variables—the ideas—in people’s minds, whatever psychologists like to think. The number of possible ideas that a single mind could hold is far greater than the number of people on Earth, so it doesn’t matter how large the sample—it won’t be enough. Moreover, it is impossible in principle to control for all these variables, because if one did manage to set things up in such a way that there were no confounding variables (this is pie in the sky, as I have indicated—see post 3, this series), one would then not be looking at what one wanted to look at in the first place, since it would be a completely artificial laboratory environment. Educational theory can never be science; it is philosophical, and none the worse for that. Some of our most important ideas are philosophical.

As Thomas Szasz says, people grossly underestimate the complexity of the human condition.

Educational theory: science or philosophy? Post 5: You can’t measure knowledge growth, or coercion, or rationality, or…

Various posters have complained that the theory that coercion diminishes creativity is testable and scientific and that in saying that it is not, I am merely trying to immunise it from empirical criticism that in fact refutes it. This is false. Such individuals tend to want a “scientific basis” for everything. This is a bit like John Gray saying that he could not find a basis for liberalism so he is going to go back to being an authoritarian conservative or something. There isn’t one. Liberalism isn’t a scientific theory either. It is not testable, not refutable; it can’t be derived from anything. The arguments about liberalism are philosophical ones, and we live in an era when philosophy is discredited, so that any philosophical argument is either rejected out of hand or interpreted as a tacky scientific theory. Kolya Wolf calls this mistake the “if it can’t be measured it doesn’t exist” idea.

Educational theory is not scientific, any more than Popper’s theory itself is. In response to this comment of mine, someone said:

“You could do an experiment. If you say knowledge grows under circumstances X but not under Y, you could look at cases where people have been irrational and had entrenched theories and see whether knowledge has grown, and then look at cases where they haven’t, and see whether it has grown in those cases. If you find that in all the cases where people had entrenched theories, knowledge has grown, and not vice versa, then you would refute Popper’s theory.”

But that is not true. Because first of all you can’t measure whether a theory is entrenched or not, and secondly, you can’t measure whether it is true or not. Perhaps my critic might reply that you can measure it “approximately according to your best guess,” but measuring it approximately according to your best guess won’t do. Because then you are not measuring whether the theory is true, you are measuring whether you agree with it. And Popper’s theory says nothing about whether successive theories will agree with you or not.

And as for entrenchment: you can’t distinguish entrenchment even approximately from someone clinging to a theory for a million reasons which are perfectly legitimate in the Popperian scheme, such as that there is a genuine problem with the competing theory that is invisible to you.

So you can’t distinguish entrenchment from non-entrenchment. You can’t distinguish coercion from not-coercion; you can’t distinguish truth from falsehood; you can’t measure any of these things. But you can argue about them and you can reach conclusions about them, from some of the most powerful arguments human beings have ever produced, namely, philosophical arguments about knowledge and so on, which culminate with Popper. (See other posts in this series [above].)

See also:

Sarah Fitz-Claridge, 1996, ‘Educational theory: science or philosophy?’, https://takingchildrenseriously.com/educational-theory-science-or-philosophy/