Radix
Carl Hendrick has published an essay on reading comprehension, and the title carries the whole argument: it is not a skill.
Schools have taught it as one for forty years. Find the main idea. Make an inference. Summarise. Predict. These were assembled by watching what weak readers fail to do and then handing the failure back as a curriculum — the symptom, written up and prescribed as the cure.
Then the arithmetic.
Measured against standardised tests rather than the instruments the researchers wrote for themselves, strategy instruction returns an effect of about 0.09. Metacognition returns nothing detectable at all. Willingham finds the entire benefit arrives inside roughly ten hours; quadruple the teaching after that and nothing further happens. There is no long apprenticeship in inference, because inference was never the thing being learned.
What was being learned is words, and what stands behind them. Laufer's threshold: a reader needs 95 to 98 percent of the words on the page already in hand before understanding is comfortable. Spencer, Quinn and Wagner tested four hundred and twenty-five thousand children and found that among those who could decode and knew the vocabulary, fewer than one in a hundred comprehended poorly. And of every intervention studied, content instruction is the only one where more time yields more effect — the signature of something being built rather than briefly rehearsed.
That sets the real difficulty, and Hendrick does not flinch from it. Direct vocabulary teaching manages three or four hundred words in a year against a requirement of tens of thousands, and taught lists do not generalise to untaught words. A thimble, carried to a reservoir.
His answer is morphology. Nagy and Anderson put it near sixty percent: that many of the unfamiliar words a student meets have meanings recoverable from their parts. Give a child the root sect, to cut, with the prefixes and endings that fasten to it, and the child does not receive five words — bisect, dissect, insect, intersection, insecticide. The child receives the rule that makes them, and the next hundred besides. Morrisey's Word Mapping Project reports four months of growth on words nobody taught. That is the difference between stocking a mind and equipping one.
The last section is the one that ought to sting. Thirty million words of gap by age three; three thousand root words of gap by the second grade. Coleman's principle, that for a disadvantaged child school time is the only academic learning time. A knowledge-rich curriculum is therefore compensatory by construction — it gives most where the home gave least. Strategy instruction, being cheap and portable and empty of content, gives least precisely where it is needed most.
Hendrick's sentence, which is the essay in nine words: "the strategy is downstream of the knowledge, not upstream."
Read it.
HE-2 believes I have failed to convey the key message of this article. He says it is, "There are many reasons to learn lots of words, but one of them is that it will make you better at using AI. Your vocabulary can be a superpower."
Nihil in intellectu quod non prius in memoria.