I’ve been reading Piotr Wozniak’s writing on learning recently. Piotr Wozniak is a computer scientist who invented the spaced repetition algorithm software SuperMemo and who writes about self-directed learning and related topics at a wiki of great depth and breadth. I have been diving into it and it is an interesting exploration of what he calls free learning. The main article that lots of the wiki pages are collated under is titled “Why I wouldn’t send my child to school”, which could easily sum up this substack too. Throughout this piece I will be referring to research and observations from across his website.
My partner and I are sitting drinking coffee in the lounge. It’s a Saturday, which always changes the pace of the conversation, at least until the second cup kicks in. In the morning the slow ogre-like thud of eldest child walking down the stairs is distinct from the more syncopated thud of youngest child shuffling down on their bum. I knew something was coming before the first beat sounded as the dog had lifted their head off my lap with sudden interest, with a look reminiscent of the way both my children look when they hear the first tinkling of the music of the ice cream van.
This is slow metronomic thudding for sure. I look over my cup of coffee and I am faced with the dog’s tail, which is anything but slow right now, going ten to the dozen. Through the door she comes and they embrace, as they do every morning, and the pace of life then picks up dramatically.
The weight of eldest child climbing on to your lap is distinct from the weight of youngest child, especially as youngest child almost always wants to simply wriggle their way into a cuddle, but here eldest child is solely reaching for the bookcase behind me; foot on groin, knee in nose, coffee on lap, there is no warning that this approach is not friendly. On complaint it turns out this is friendly fire, a pen that is in one of the drawers is what is required and communication beforehand, apparently, is not.
Two sheets of yellow paper are produced from another drawer, thankfully on the other side of the room, and writing ensues.
“How do you spell going?” she asks. And five minutes into her day she is self-directing her learning.
Five minutes later she is licking her own chin, in fact we are alternating between licking our chins and trying to eat them with our top teeth, and we are laughing heartily and learning how to read at the same time.
I am going to describe a scene that I imagine could play out in many homes across the country on a weekend morning. But as always the devil is in the detail. Or rather, I am going to argue that it needn’t be. Children are designed to learn, we have decided to put the devil into the detail of educating children. But the detail can be and should be quite simple.
The previous day we had been to the pet store on a whim whilst shopping for food. Whilst cuddling the dog my daughter had decided that it might be a fun to head back to the pet store and this time take the dog. But then, she had reasoned, it is also a treat to go on a nice walk, especially with my brother-in-law’s border collie. Why not let the dog decide she thought. The pen and paper had been to write the two options down. Going to the pet store. Going on a walk with your friends.
One of my roles in life is begrudging assistant. And so when asked to write these notes for her, I obliged but in the character of a man who had forgotten how to spell some parts of the words. I couldn’t remember the “th” for the. But with her assistance we made it through scribing going to the pet store.
When asked to write the second note it occurred to me that there was an opportunity to forget how to spell friends, and discuss the difference between the “th” sound and the “f” sound in English, and hence how we ended up with some over-exaggerated examples of how to make those sounds with our tongues and top teeth all over our chins laughing at each other.
Piotr Wozniak has a fundamental law of learning. He believes that ‘good learning is inherently pleasurable, and without pleasure there is no good learning,’ and this definitely was pleasurable. She then proceeded to tell us her plan. The pieces of paper would be folded with a dog treat inside and the dog could then choose which of the two adventures they would go on. Here the dog is with her chosen treat.
From this initial idea, we were able to do not only some phonics, but some paper folding techniques; some discussions on how to lay the two options out to remove any possibility of bias, hinting at the scientific method; some basic dog psychology; and when our dog choose the treat wrapped inside a walk with her friends we were able to discuss what it means to be a dog owner, what are our responsibilities and relationships to our pets, which led to my daughter joining me on an hour’s walk around the park where she got to go on the climbing frame and start her day with some physical activity.
There is something to be said about how one simple idea can spawn a lot of different activities out of it, and that unschooling has this flexibility and responsiveness is a definite advantage over school, but I want to focus in on something more particular today.
I’ll say it again: this is a scene that could have played out in many homes that Saturday morning. In fact I honestly believe that there were probably many young girls her age who walked their dog, climbed in the park and did some phonics that morning. There was also, almost inevitably, some who made dog treat holders or wrote notes to their dog too. This behaviour is not atypical for an imaginative seven year old with a pet. But what is atypical, and worth noting I think, is why her five minutes of sounding out these fricative phonemes is an unusual cognitive experience for a seven year old.
But before we can get to that point we need to back up and walk through some of Wozniak’s ideas on learning and the terms that he likes to use. For what we have been calling unschooling or self-directed education Wozniak prefers the term free learning, learning rooted in freedom, where ‘there are no constraints of time, space, pace, method, direction, or subject matter.’ The reason that this is preferable to school is because it is evolutionary adaptive and will lead to compliance with his fundamental law of learning. Humans are naturally curious, we know this of babies, toddlers, and seven year olds with ideas of dog treats buried inside notes and promises to pets. Yet curiosity has multiple slightly different meanings, some carrying negative consequences, so Wozniak prefers the term learn drive.
The learn drive is a natural tendency of the brain to seek new information. The learn drive is generated by a guidance system hard-wired in the brain. This guidance system activates the reward centers upon detecting novel patterns in memory storage.
Learn drive is a vital evolutionary adaptation and employs information entropy to make judgements about the environment and internal inputs.
In 1948 Claude Shannon proposed the concept of information entropy and since then neuroscientists now know that the brain can indeed detect signals with different levels of entropy delivered, via the senses, to the hippocampus.
But for Wozniak there is a subtle difference between information and meaning; information entropy encodes the information but not the meaning. Or to put it another way each brain responds to information differently based on prior knowledge. A Dutch news broadcast has the exact same information but a different meaning depending on whether you speak Dutch or not. So we need to derive a new term that he calls learntropy, which can be seen as measure of attractiveness for a particular brain.
There is no optimum entropy level for a channel. There is only an optimum entropy level that fits a specific brain.
Here we can briefly pause and return to my daughter and see that our interaction in the midst of her desire to provide our family dog with an interactive puzzle probably had a high level of learntropy. The small amount of information encoded, the personal nature it was delivered, by a trusting person responsive to her behaviour and enjoyment levels and prior knowledge of phonics kept the signal not only high in information but also meaning.
The brain uses semantic learning, associating new knowledge with old knowledge, and this builds memory coherence. That I know as well as anyone where she is at on her reading journey means that the information I impart has high learntropy and therefore a greater likelihood at building coherent memories.
Great. But what about the other imagined seven year olds also engaging in similar activities on a Saturday morning. Why is this different compared to them?
Well here we can turn to Wozniak’s thoughts on memories and how pleasure and learntropy affect how the brain makes memories.
He tells the story of an eight year old Polish boy he knew whose mother was helping him learn English with Wozniak’s spaced repetition software. The boy had two different templates, one for school vocabulary and one template for things that he liked. The two templates had different colours. The mother, when inputting the English word sausages, accidentally duplicated the item, but also managed to put one item in the English word template and one in Thomas’s preferred template. In effect she created two identical items with two different templates: the liked template and the disliked template.
Over the next three months the boy failed the item in the school template eleven times but apart from failing it on the first attempt he correctly recalled the item in his liked template ten times over the same period. The colour, and the association that Thomas had with school, literally changed his ability to remember the word. On average he viewed the word every eleven days.
But another event takes the point further. Four months after first placing the word sausages in her son’s spaced repetition software his mother bemoaned the fact that he couldn’t remember simple English words to a friend over dinner and asked Thomas what they were eating. He replied instantly, “frankfurters.” She was shocked. How did he even know this word. It turned out his grandfather who had visited almost a month ago and cooked frankfurters on the bbq had asked Thomas “Do you want some frankfurters?”
‘One memorable event may count for more than dozens of repetitions at school’
Herein lies the atypicality. It is not that as a parent I might spend some time getting my child to think and not fink on a Saturday morning, but that my daughter’s relationship to phonemes is one of sparseness compared to a child her age attending school and yet each and every individual event, every interaction with phonics, is memorable. (They are memorable precisely because of this infrequency, or more to the point, because she dictates the frequency and this infrequency, at the moment, suits her.)
We know memories are consolidated whilst sleeping, but before memories can be consolidated they need to be valued first, and this is why pleasure is key to learning.
Wozniak believes that displeasure in learning carries a risk of toxic memories, as the brain cuts off the consolidation signal and the recall signal as it considers them of low value. The colour associated with the school template reduced the value of the word sausages compared to the liked template. And here is the devil. Lying in the details of the thing that we call education: schools enforcing prescribed curriculum upon young people in an authoritarian environment.
School is a place of coercion and coercion is not likely to provide young people with much pleasure. The learn drive is being overriden, replaced with the school drive, and memories are not cohering, nor consolidating. Instead, Wozniak argues, school inculcates learned helplessness in young people as schools ‘incentivize the transition from failure to harder failure’. Wozniak’s whole oeuvre is polemical stuff indeed and well worth exploring.
The great sadness is not only the lack of freedom children have to endure, but the great waste of time that all this takes. Herein lies the somewhat unique cognitive experience that my daughter, and other free learners, have. Her “reading comprehension” is without doubt below most children her age, however, the amount of time she has spent on drilling phonics is insignificant in comparison. I have already written about how unschoolers learn to read at wildly different ages and by the time she is reading at the same level as her peers she will still have spent significantly less time on phonics, but she will still have her learn drive intact and she will probably be a passionate reader, as most self-directed learners are when they finally get around to reading books on their own, no matter how long, or how beautifully messy, the journey is.
Yes, it is true that unschoolers will not know everything on the curriculum, but it is also true that everyone who goes to school won’t either. But that is not the point. It is absurd to want everyone to know exactly the same stuff anyway. Maintaining a healthy learn drive should be the core principle guiding every educational decision.
A child with a healthy learn drive will jump from thing to thing building semantic knowledge as they go as the things that they are jumping to and from will invariably be of great interest to them as they will all be freely chosen. In the process of undertaking that sporadic and unique journey they will inevitably pick up core skills such as reading through such odd and idiosyncratic ways as wanting to write notes and conduct experiments with their dog. It’s not, as the diagram shows, about getting from A to B but wandering all over the map learning about the territory called life as you go.