My friend M. has been deep in the books, trying to design and redesign a good homeschooling environment for her eight-year-old son who is diagnosed with autism spectrum disorder, and who really needs quite intense attention. She is looking at various styles of teaching and how they are shown to work with real children.
She came to me a few days ago saying (I'm paraphrasing): "You know what's really amazing about the so-called 'classical model of homeschooling?' So many homeschoolers put so much stock in it -- in this idea that the medieval trivium corresponds to three stages of child development -- but there isn't really any evidence for it at all! If you go tracing back the claims that people make about classical homeschooling, they all go back to just one source --"
"'The Lost Tools of Learning,'" I said. "The essay by Dorothy Sayers."
"...And, well, she just made it up! She doesn't claim to know anything about how children develop. It's just her idea. But this entire industry has sprung up to support classical homeschooling, as if there was any evidence that designing a school the way she imagines it is something that really produces good outcomes. There isn't any evidence. And of course the medieval trivium wasn't at all like the homeschooling trivium that people talk about. There's no evidence that children ever were taught the way she describes."
I thought for a moment how to respond. "I think that the essay provides a useful organizing principle," I ventured. "No more than that." I did have to defend myself a little bit!
And I do think that the organizing principle behind "The Lost Tools of Learning" is so useful that I maintain it's part of the "homeschooler's canon" of educational philosophy.
+ + +
But M.'s reminder is useful, too -- that Sayers' essay is just that: an opinion piece. After all, I do go around telling other homeschoolers, when I need to, that I am "oh, sort of a classical homeschooler, with a little Charlotte Mason thrown in." Really, this is just a piece of code, shorthand for what I am not: I am not an unschooler, I am not a school-in-a-box-er, I am not entirely disorganized (I have some kind of plan), and, well, I'm not entirely "evidence-based."
(Even the Charlotte Mason reference is really cheating. I've never actually read any serious books by or about Charlotte Mason. That reference is code for "I like making my kids tell back what they've learned, and I tried keeping a nature journal for a while.")
One careful reading of Sayers' essay will show you that she never intended her imaginary school to be set up as a real place. She breezily assumes ideal and impossible conditions:
Let us amuse ourselves by imagining that such progressive retrogression is possible. Let us make a clean sweep of all educational authorities, and furnish ourselves with a nice little school of boys and girls whom we may experimentally equip for the intellectual conflict along lines chosen by ourselves. We will endow them with exceptionally docile parents; we will staff our school with teachers who are themselves perfectly familiar with the aims and methods of the Trivium; we will have our building and staff large enough to allow our classes to be small enough for adequate handling; and we will postulate a Board of Examiners willing and qualified to test the products we turn out. Thus prepared, we will attempt to sketch out a syllabus--a modern Trivium "with modifications" and we will see where we get to.
But first: what age shall the children be? Well, if one is to educate them on novel lines, it will be better that they should have nothing to unlearn; besides, one cannot begin a good thing too early, and the Trivium is by its nature not learning, but a preparation for learning. We will, therefore, "catch 'em young," requiring of our pupils only that they shall be able to read, write, and cipher.
Did you catch that bit -- that at the start of her "school" the children have already been taught to read, write, and do arithmetic? We're on our own for that part, homeschoolers.
Sayers' imaginary school is not, actually, a plan (and she takes pain to point this out). Nor is it a reconstruction of the medieval trivium in any way -- significantly, she stresses, "It does not matter, for the moment, whether it [the trivium] was devised for small children or for older students, or how long people were supposed to take over it."
No, this school she describes is a rhetorical device. The point of the description is to create vivid pictures in the readers' minds, of children arguing, or finding Cassiopeia in the night sky, or examining portraits of the Kings of England, or carefully studying maps. This is just an illustration to motivate readers to hear and accept her philosophy of education, which she emphasizes by making it the very last sentence of the essay:
For the sole true end of education is simply this: to teach men how to learn for themselves; and whatever instruction fails to do this is effort spent in vain.
This is a philosophy that appeals to me, and it is why I go around telling people that I am a "classical homeschooler." Not because I think my homeschool should be split up into Grammar and Rhetoric and Dialectic (although I sometimes use those terms to describe the level of mental process that is engaged by a particular book or curriculum). I tell people I am a "classical homeschooler" because I believe my job is to teach my kids how to teach themselves.
+ + +
Which brings me to the problem of evidence-based teaching. Evidence is overrated.
If I say "Evidence is overrated," you should know I am speaking tongue in cheek. I am trained as an engineer and as a scientist, and even though I abandoned that life in its early stages, its inner life has not abandoned me: I can never lose the engineer's problem-attacking method, nor the scientist's reflexes to observe, test, and revise.
No, what I really mean is that the marshalled evidence often overemphasizes one kind of data to the exclusion of another kind of data that, depending on the application, may be more important. The distinction between the two kinds of data: that drawn from large groups and that drawn from individuals.
Data drawn from large groups can be usefully applied to help individuals, it's true. Let's imagine an experiemental school -- would that schools would actually do this -- in which two methods for teaching arithmetic are to be evaluated. Children are randomly assigned to two classrooms, one in which the arithmetic skills are taught via method A, and one in which the teacher uses method B to teach the same skills. After a suitable time, the children are evaluated, and method A comes out ahead: perhaps more students in classroom A have achieved a certain minimum competency, or perhaps the average score of students in that classroom is significantly higher than the average in classroom B. Perhaps A outperforms B by both metrics. If the test is repeated over different schools and different populations, and if A consistently outperforms B, that would certainly be a strong argument for teaching with method A instead of method B. I would gladly support this kind of approach to curriculum selection in my local public schools, who are, after all, in the business of educating large groups.
However, just because A is the better method overall does not mean it is the better method for every child in the class. Imagine there is one particular child -- let's call him Jacob -- who responds better, in a critical way, to method B. It's not that Jacob was one of the top kids in classroom B -- no, let's be honest, most of the "top kids" in classroom B would also have been "top" in classroom A, because every classroom contains at least a few kids who are going to absorb material no matter how you throw it at them. No, this boy was on the margins: Jacob was one of those who achieved competency in classroom B, did well enough. But when the classrooms all switched to the A method -- the method that was proved superior by the average-score metric and by the number-achieving-minimum-competency metric -- Jacob did not succeed, did not achieve competency. More of his classmates will succeed in the A environment, and that is good for them, and good for the school, and good for the teachers, probably. It does not help him.
I am not, mind you, saying that the evidence drawn from large groups is unhelpful for the large groups. I do think schools should use data to shape their curriculum, because that is the job of the schools. I just want to point out that the schools have one job -- to teach the large groups -- and Jacob's parents have another -- to teach Jacob.
This is, to me, the distinction that gives meaning to the notion that education of children is primarily the responsibility of their parents. It is reflected in Minnesota law, at least: Our state's constitution stipulates that the legislature shall establish a "general and uniform" system of public schools; "uniform" is supposed to provide for equal and fair access to schooling, and that is commendable and appropriate for government schools. The statutes specify this: "The parent of a child is primarily responsible for assuring that the child acquires knowledge and skills that are essential for effective citizenship."
My point: You use large-group evidence to design approaches that are aimed at large groups. If your job is to teach Classroom B, you use what you've learned from metrics that measure Classroom B. But if your job is to teach Jacob -- and that, mind you, is not legally our state's job, nor a single classroom teacher's job, nor can it be -- it is Jacob's parents' job -- then you have to give primary weight to evidence drawn from Jacob and not to that drawn from his classmates.
+ + +
I derive a lot of inspiration from Dorothy Sayers' opinion piece, even though it is not evidence-based. Sayers never thought that people would try to copy the imaginary school she described, and I don't either. I follow her example -- not by trying to create in my home the school that she imagined -- but by trying to create in my home the school that I imagine.
I consider the problem that life has posed me -- and I frame it and sketch it out as How will I teach my children to teach themselves? I make assumptions, I list my constraints, I weigh costs against likely benefits, I approximate when necessary, and in the end I do what I can with what I have.
I make a guess -- this particular sort of curriculum will interest my eleven-year-old, or arranging the school desks in that particular way will keep my eight- and five-year-olds from fighting; I try it out; I check to see if my eleven-year-old is engaged in his work, I listen to see if the squabbles have diminished; I revise my hypothesis, I try again.
+ + +
There is another thing to weigh, too -- and it is sort of outside the evidence of what sort of methods produce academic excellence. Because academic achievement is not, actually, central to what I am trying to do here (although it's very easy to forget that, because academic achievement is so much more easily measured, and because the culture values it, and I am swayed by culture).
What I am supposed to be doing is forming human beings who are honest, fair, kind, and wise, who have a firm understanding of their own value and of the value of other human persons, who recognize their direct responsibilities and carry them out. And no amount of evidence of academic achievement should entice me to apply a method that undermines any of these.
Meanwhile, I am supposed to be forming myself in honesty, fairness, kindness, and wisdom -- and if I somehow form my children in ways that deteriorate my own character, I'm equally wrong.
To put it bluntly, if by daily beatings or regular bribes I could ensure my child a perfect SAT score, that wouldn't mean I should do it.
+ + +
Large-group evidence based on large-group metrics will only take you so far. It isn't useless: it's just that it is useful for purposes different from mine. (I can, of course, use it for my own purposes: what's good for the group may be the best thing to try first, before tweaking to suit the individual.) This is true about education, but lots of other things too: taking care of your health, disciplining children, deciding what sort of community to live in, deciding whom to spend time with. Sometimes "what works" is what matters; other times it's far more important to find something you can live with, whether it "works" or not; most often it is a matter of balancing both.