A remark by @autumnpard on Memorization reminded me of an analogy I came up with some time back: A song by Jakob Dylan includes the following lines: Cupid, don’t draw back your bow Sam Cooke didn’t know what I know …note that in order to understand these two simple lines, you’d have to know several things:
1) You need to know that, in mythology, Cupid symbolizes love
2) And that Cupid’s chosen instrument is the bow and arrow
3) Also that there was a singer/songwriter named Sam Cooke
4) And that he had a song called “Cupid, draw back your bow.”
“Progressive” educators insist that students should be taught “thinking skills” as opposed to memorization, and the advent of LLMs has further driven such thinking But consider: If it’s not possible to understand a couple of lines from a popular song without knowing by heart the references to which it alludes–without memorizing them–what chance is there for understanding medieval history, or modern physics, without having a ready grasp of the topics which these disciplines reference?
And also consider: what’s important is not just what you need to know to appreciate the song. It’s what Dylan needed to know to create it in the first place. At least in theory someone who heard the song and didn’t understand the allusions could have spent 5 minutes googling and figured them out, although this approach wouldn’t be exactly conducive to aesthetic appreciation. But had Dylan not already had the reference points–Cupid, the bow and arrow, the Sam Cooke song–in his head, there’s no way he would have been able to create his own lines. The idea that he could have just “looked them up,” which educators often suggest is the way to deal with factual knowledge, would be ludicrous in this context. And it would also be ludicrous in the context of creating new ideas about history or physics.
To use a computer analogy, the things you know aren’t just data–they’re part of the program. I’ve seen no evidence that there exists a known body of “thinking skills” so powerful that they bypass the need for detailed, substantive knowledge within specific disciplines. And if such meta-level thinking skills were to be developed, I suspect that the last place to find them would be in university Education departments.
There are skills which facilitate thinking across a wide range of disciplines: such things as formal logic, probability & statistics, and an understanding of the scientific method–and, most importantly, excellent reading skills. But things like these certainly don’t seem to be what the educators are referring to when they talk about “thinking skills.” What many of them seem to have in mind is more of a kind of verbal mush that leaves the student with nothing to build on.
There’s no substitute for actual knowledge. The flip response “he can always look it up” is irresponsible and ignores the way that human intellectual activity actually works.
None of which is to say that traditional teaching practices were all good. There was probably too much emphasis on rote memorization devoid of context–in history, dates soon to be forgotten, in physics, formulae without proper understanding of their meaning and applicability. (Dylan needed to know about Sam Cooke’s song; he didn’t need to know the precise date on which it was written or first sung.) But the cure is to provide the context, not to throw out facts and knowledge altogether–which is what all too many educators seem eager to do.
There really does seem to be a deep-seated hostility toward knowledge itself among many who define themselves as “educators.” And a lot of students today are all too eager to use LLMs to do all of the work…or as much of it as they can get away with…to guard themselves against either learning anything at all or developing the ability to do focused and concentrated work.
See my earlier Thinking and Memorizing post, also Classics and the Public Sphere.
Your thoughts?
Bingo: we need substantive knowledge in specific domains, from which we can generalize, analogize, compare/contrast. Facts without theory are sterile, but theory without facts is baloney.
Thanks for stating this so well.
I had for the sixth grade, a marvelous, creative and demanding teacher – Mr. Terranova. I honestly believe now that he was operating under a lot of Montessori theories – but one of his practices was having the classes memorize poetry. We had to memorize a poem or a stanza of a poem every night and recite the next day in class, and some of them were quite long. Robert Service’s The Cremation of Sam Magee was one of them. (Also the preamble to the Declaration of Independence…) There exists a theory concerning brain development (I have read somewhere or other) that posits that the brains of children of a certain age are like sponges and need to be filled, and memorizing materiel like poems and the multiplication tables is perfect for this – especially when it is useful information. The sheer discipline of memorizing something is useful in the long run, too.
I am very much in favor of your position. We all, especially children, need to acquire “knowledge” that we can put in context and develop understanding. Col Baud recently reminded Daniel Davis that intelligence refers to “understanding” something, in this particular case “the enemy”. A most important point. The key point is to proceed from knowledge to understanding and get to “wisdom” in how to apply the knowledge we have come to understand. Among those studying “wisdom”, there are not only the propositional and perspectival elements but also procedural and participatory ones involved. All involved need to be aware that this not a “one and done” but a commitment to a worthy goal of cultivating a way of life. This lesson has been ignored for far too long.
Yeah, there is room to joke or to seriously question whether I have any thinking skills.
But, if I in fact do, lots of getting there was sponging, memorization, working on observation which can be closely tied to memory, and stuff in the traditional liberal arts of logic, rhetoric, grammar, harmonics, geometry, arithmetic, and astronomy. My grasp of astronomy is extremely weak, so that maybe does not count. (I struggled with geometry, and my formal grounding in the first three are also shallow.)
If I can think, my first guess is that many educators cannot. Because I tend to define tasks easy or possible for me as tasks that are easy for everyone, and assume that my ability is the baseline to hold everyone to. (If there were just one area where I was really strong, this would obviously be verifiably untrue.)
Given the questions some Education majors ask, we can probably infer that the foundation for their research is unsound. Thus, Education majors perhaps do not have a place a research universities, and perhaps should not be directing ‘improvements’ at the university level. However, the path for inferring that is understanding the industrial engineering arts for quality control in mass production, and the basis those have in what electrical engineers or mechanical engineers know about manufacturing of individual parts and how those relate to the progression of defects in individual parts to failure. So there is some basic statistical reasoning that students are probably /not/ learning in highschool. Or, at least I did not learn it in highschool, I learned it on the job, and as a hobby, mostly outside of schooling.
Anyway, if the education majors did not learn this in primary, secondary, and tertiary school, it is a little unfair of me to blame them for not knowing it. Unless the people who taught them were derelict in passing them off as trained in a tertiary school.
At a very significant level, I do not know what we think tertiary schooling is, or should be.
(Given the race war conspiracy questions that some education majors ask, we need a very fine and individual model of student learning to capture all variables. The experimental tools of the education major are very likely crude and noisy in comparison. The simplest answer is that Education research is invalid, and that Marxist idiots being credentialed as educators is the most unjust thing that society is doing to primary and secondary students.)
That aside, what is thinking?
One notion is that an educated person is equipped to be paranoid, to systemically distrust every academic field on its own individual merits, and to trust no research that they did not personal do themselves. That an educated thinker may know what everyone says for the broad strokes of many fields, but has no reason to repeat those statements, because they have confidence in their own thought and training, and need not borrow any weight or authority that comes from outside. This is an approach that synergizes well with treating most ideas as related, and also testable in groups and individually. Ideas are more respectable if they pass tests, tests applied impartially and ‘without favor’; so that a thinker will punish even an idea that they like and prefer if the testing indicates that it is untrue.
A competing model is that ideas are strictly individual, and can be treated as entirely uncoupled, not matter what one is purporting to say about complex interrelationships. Authority must come from outside the individual, and what matters is a fast update time, and never being caught stating an individual idea that is not the currently political consensus expedient view of that claim.
Another competing model is that education is strictly style of speech, and selection of words. Someone with the correct style is educated, even if they way they use ideas shows them to be a henwit. Someone without the correct style is uneducated, even if they show some art with their ideas. This model is particularly a problem with henwits, especially ones who have misrecorded economic values, think that they can just murder the ‘uneducated’, and things will work out great for them personally and economically. It means that credentialled ‘elites’ whose business is in acting as proxies for others may have directly undermined the future of that business.
Talk of Artificial Intelligence is the rage these days with many thinking it is a panacea and will ruin mankind. But someone gave me a great analogy on its limitations.
Suppose you asked it to design the most fuel efficient car – and the only thing in its memory was carburation – not electronic fuel injection.
So in the best case scenario it would come of with the most fuel efficient carbureted car it could.
It would have been too much of a leap to have it design an electronic fuel injection system, for which it had no reference data.
All of us are ignorant, just about different things. One of the insufferable aspects of English class snobbery has always been based on classical allusions and imagery. For the upper classes, having a positive aversion to anything that might considered “trade”, an encyclopedic knowledge of Roman and Greek mythology was as far from anything useful as they could desire. One of the things preventing me from pursuing poetry is the habit of layering all the “good” ones with now obscure references. I’ll also note that Shakespeare managed a quite successful career and legacy in a much less literate age than today without shying away from the classics, it’s all about context.
Ironically, the classical reference was clear, even to me, while a reference by a song writer I’ve never heard of to another that Is just, barely, familiar doesn’t really make a compelling argument for me. I can pretty much guarantee that most allusions to pop music will pass cleanly over my head, yet I dare to hold up my head in polite company.
I also understand that the question under discussion is what do we all need to “know” for society to function. What’s the lowest common denominator? At the risk of throwing out an allusion of my own that may be obscure to some. (I hope not.) Of the limited number of hours available for universal education; how many should be devoted to Bullfinch’s Mythology and how many to Young’s modulus?
I’ll throw reading and math, up to algebra I into the pot. (Another of those allusions.) Feel free to raise.
The classical learning model, often used in home schooling, is the trivium which relies on a three layered approach often depicted as a pyramid
The bottom layer is grammar (knowledge) providing a basic framework of facts and skills from which knowledge can be built
The next layer is logic (reasoning)for critical thinking, reasoning and the ability to analyze information
The third and final layer of progression is rhetoric – communication and persuasion skills to express ideas
Susan Wise Bauer in her books uses the trivium as a guide for home schooling
The third layer isn’t just about communication, but represents a final layer of mastery
The model can be used for many areas, even in the trades. An aspiring car mechanic needs to understand the basics of a car and how it fits together and works (grammar) The mechanic can then progress to being a professional by analyzing information the car provides and developing solutions for repair (reasoning) Finally there is the ability to manipulate the prior two skills to gain new knowledge or affect the large environment (rhetoric) – say by customization.
I use this basic model in forklift training
Yes memorization and basic understanding is necessary because it is not so much data and frameworks that provide fuel but it provides as the name suggests the basic grammar by which one can interpret the world.
Classical education, by fusing the empirical and the metaphysical, looks to create true knowledge by creating order out of a chaos through a framework of language and perspective. As the great comments above state, you need both theory and data.
As a side note, the problem of modern science is that it separates empiricism and the metaphysical. In fact you can argue that the past 400+ years of the modern enlightenment has been a full-scale flight from Aristotle. Science sharpens itself as a tool by limiting its perspective, it can describe but not explain.
There is another perspective using Boyd’s OODA which derived from Gödel’s Incompleteness Theorem in that understanding a given situation cannot be accomplished internally but only in conjunction with a n external framework.
Bill Brandt…”Suppose you asked it to design the most fuel efficient car – and the only thing in its memory was carburation – not electronic fuel injection” That is a very interesting question: In principle, could the LLM link together the knowledge that more complexity in the fuel algorithm can lead to more fuel efficiency…and the knowledge that a complex control algorithm can be instantiated more cost-effectively with electronics rather than with pure mechanics…and come up with an ‘invention’ of electronic fuel injection? It doesn’t seem impossible, but I don’t see how to test it since EFI is *already* known to the model.
Tried an experiment with Grok’s inventiveness skills:
I need an invention. Supermarket food carts are very irritating because if the parking lot is not absolutely level, they will run away while you’re trying to put your stuff in the car. Need an approach which will hold the cart in place but still doesn’t require additional effort from store personnel to return the cart to the pickup area. Much be cheap to manufacture and resilient & low-maintenance in service.
Invention: Magnetic Brake Cart System
Concept: Equip supermarket carts with a simple, low-cost magnetic brake system that engages automatically when the cart is stationary, preventing it from rolling on sloped parking lots. The brake disengages effortlessly when the user pushes the cart, requiring no extra effort or action. The system is durable, low-maintenance, and integrates seamlessly with existing cart designs.
How It Works:
Magnetic Brake Mechanism:
Each cart wheel contains a small, spring-loaded magnetic brake pad integrated into the wheel hub.
When the cart is stationary (no force applied), the spring pushes the brake pad against a metal disc in the wheel, creating enough friction to hold the cart in place, even on moderate slopes (up to 5-10% grade).
When the user applies minimal forward or backward force (e.g., pushing the handle), the motion overcomes the spring tension, retracting the brake pad and allowing the wheel to roll freely.
The magnetic component ensures consistent braking force without wear, as it relies on magnetic attraction rather than mechanical clamping.
User Experience:
The brake engages automatically when the cart stops, requiring no user input (e.g., no buttons or levers).
Pushing the cart feels natural, as the brake disengages with minimal effort (comparable to standard cart resistance).
No additional steps are needed to return carts to pickup areas, as the system doesn’t alter the cart’s mobility for store personnel.
Design Features:
Materials: Use neodymium magnets (small, powerful, and affordable) and durable, weather-resistant plastics for the brake pad and spring. The metal disc is a low-cost steel component embedded in the wheel.
Cost: Estimated additional cost per cart is $5-$10, as the system uses inexpensive, mass-produced components (magnets: ~$0.50 each, springs: ~$0.20, plastic housing: ~$1).
Durability: The magnetic brake is sealed within the wheel hub, protecting it from water, dirt, and debris. Neodymium magnets retain strength for decades, and the spring is designed for millions of cycles.
Maintenance: Virtually maintenance-free, as there are no complex electronics or exposed parts. Wheels can be replaced as a unit if damaged, similar to current cart maintenance.
Integration:
The system can be retrofitted to existing carts by replacing standard wheels with brake-equipped wheels.
No changes to cart frames or store infrastructure are required.
Compatible with cart corrals and stacking systems, as the brake only engages when the cart is stationary.
Advantages:
User-Friendly: No learning curve or extra effort for customers or staff.
Cheap: Low manufacturing cost makes it feasible for widespread adoption.
Resilient: Weatherproof, durable components withstand harsh outdoor conditions (rain, snow, heat).
Low-Maintenance: No regular servicing needed, reducing store costs.
Scalable: Works on any cart size or type, from small baskets to large flatbeds.
Potential Challenges and Solutions:
Challenge: Brake strength must balance holding power with ease of movement.
Solution: Calibrate spring tension and magnet strength during manufacturing to ensure the brake holds on slopes but disengages with light pushing force (tested for 5-10 kg push force).
Challenge: Debris or corrosion could affect wheel performance.
Solution: Use sealed wheel units and corrosion-resistant materials (e.g., stainless steel springs, coated magnets).
Challenge: Cost-conscious stores may resist added expense.
Solution: Highlight reduced liability from runaway carts (e.g., damage to cars or injuries) and market the system as a customer satisfaction feature.
Manufacturing Feasibility:
The system leverages existing wheel manufacturing processes, with minor modifications to include the brake assembly.
Neodymium magnets and springs are widely available and inexpensive in bulk.
Prototyping and testing can be done with 3D-printed wheel housings and off-the-shelf components.
Next Steps:
Develop a prototype wheel unit and test it on a sample cart under various slope conditions.
Conduct durability tests (e.g., 10,000 brake cycles, exposure to water/salt).
Partner with a cart manufacturer to integrate the system into production lines.
This solution addresses the runaway cart problem with a simple, elegant, and cost-effective mechanism that requires no behavioral changes from customers or staff, ensuring easy adoption and long-term reliability.
“Use neodymium magnets (small, powerful, and affordable) and durable …”
Seems Grok missed the point that Rare Earth magnets come (or used to come) from China, and nowhere else in scale. Now China has put restrictions on exports of those magnets. Grok might as well design a system based on Unobtanium.
AI is not really capable of creativity. Recognizing ways in which siloed knowledge can be brought together — Yes. But inventing something de novo — No.
Neodymium is mined found in other places as well as China, 70% comes from China but there are new/reopened mines in the US and Australia. The real bottleneck is probably mainly the processing, mostly done in China but at least 2 US companies and one Australian company are getting in the game.
This substance is too valuable for the US to do without; we’re going to have to either scale up the domestic & friendly sources or capitulate to China to the level that we can obtain it from them. The first alternative is a lot better.
Grok misses that there has to be a linkage between the operator and the brake mechanism otherwise, there’s no way for the brake to differentiate the force of being pushed from the force of being on an inclined plane. As it is, it is a peculiar sort of perpetual motion machine. Assuming a force of 20kg (44lb) releases the brake; what engages the brake? Maybe a few of the people I see at the store could do it without a running start. It would sure change the shopping experience.
Considering the trouble that most places have with providing carts that have four wheels that both turn and point in the same direction, this would solve the problem of runaway carts by insuring that none of them moved at all.
The “solutions” are all assertions that run 180 degrees from everything I’ve learned in the last too many years. The first is that adding an intricate mechanism to something that costs less than $10 and now has either one or two moving parts can be done for low cost. Seals leak and present considerable cost of production, etc. Anybody that disagrees is welcome to prove me wrong, you can wave the first dollar you make from it under my nose and I’ll be amazed. Better yet, tie it to a rock and drop it from your private jet as you fly over.
Until then, I’ll keep parking so I can nudge the basket against my bumper or tire.
If Grok wants to impress me, give me drawings, a bill of material with costs, including assembly, projected failure rate and maintenance cost. Oh, a patent search that shows some misguided soul hasn’t already wasted years of his life.
The force from being on an inclined plane depends on how steep the inclination is, of course. If we stipulate that the parking lot has a 4 degree inclination and the cart+contents weighs 50 pounds, then the run-away force will be 50*sin(4 degrees), which is 3.5 pounds.
I’d judge the force needed to move a cart with operable wheels at around 5lb. I could take a dynamometer with me and see, they probably wouldn’t look at me any more strangely.
A really simple answer that seemed to elude Grok is to just put a brake caster, the kind with a lever that you kick, on one of the fixed wheels. Total cost for that size and capacity: about $2. Not prohibitive on something that costs a few hundred each and I’d gladly trade that for the seat belts that haven’t been used ever. How long it would take the average consumer to catch on, I leave as an exercise for the reader.
I can imagine a fairly simple mechanical linkage where the handle bar would move slightly when pushed or pulled to release a spring brake. Considering the abuse these carts take, I can also foresee keeping it working to be non trivial. The alternative is a few signs disclaiming liability in the parking lot and their commercial liability insurance. Buying lawyers by the dozen.
As I said, I’m sure Grok knows a lot more about obscure song lyrics than I do, but I’m satisfied I know a lot more about mechanical design then it does. For now.
AI seems somewhat OK at providing an answer if it can pull it from the internet somewhere. It’s not as if there isn’t an extensive literature covering brakes, all the relationships are well established. As always, it’s telling a good design from a problem.
Speaking of which, in the last few days a series of GM recalls has hit the interwebs. Where it stands now, GM will be replacing around 900,000 6.2l gas engines, at least. Avoiding that sort of thing would be worth a lot more than $20 a month if AI could do it.
Sgt. Mom.
I had for the sixth grade, a marvelous, creative and demanding teacher – Mr. Terranova. I honestly believe now that he was operating under a lot of Montessori theories – but one of his practices was having the classes memorize poetry.
I once substituted in a fifth grade class where the teacher had the class memorize a lot of poetry. Every student stood up and recited a poem–each student with a different poem. Each student had a moment in the sun where the entire class was listening to the student. As each student wanted to be listened to, that provided incentive to be quiet when other students were reciting. A virtuous circle.
I was told that this was a class of “problem students.” Memorizing worked for them. They learned something. They were given an audience, and similarly learned that if you want to be listened to, you need to listen to others.
If you don’t know facts, you have nothing to apply a theory to.
David F.: “This substance [neodymium] is too valuable for the US to do without …”
You are, of course, correct. The issue as I understand it is that “Rare Earths” are actually not that rare; the problem is that they occur in low concentrations and extracting them is the kind of nasty industrial process which no self-respecting Far Left Greenie in the US or Europe will tolerate on their home patch. They want the magnets, but they want someone else far away to suffer the downsides of producing them. The analogy with Colonialism is obvious, but that apparently does not bother the Greenie.
Look on the bright side! When the US Deep State forces us into an unnecessary war with China, presumably Chinese newspapers will have a lot of fun mocking the US for taking neodymium magnets off of supermarket trolleys to build weapons. Computer chips and washing machines — what goes around comes around!
Your local wind turbine actually has tons of rare earth magnets in each, the shopping carts are safe.
I’m betting China will miss 30-40% of their food supply before we miss a few magnets.
Speaking of rare earths:
https://quillette.com/2025/05/12/the-rare-earth-wars-china-trump-taiwan/
Late news just before I went to bed, trade peace in our time:
https://nypost.com/2025/05/12/us-news/us-and-china-reach-deal-to-slash-tariffs-officials-say/
Next will be explanations why it’ll work this time for sure.
I occasionally read SF, though as time goes on, less and less. SF probably represents the mother lode of notions for how “proper” socialism should work, maybe the reason I find it less and less diverting. I’m sure the idea that true AI will enable socialism “done right” appears many other places and I have just missed it. Explanations of just how great the world will be when AI is determining just how many sheets of toilet paper you need so that not one sheet more need be produced are forthcoming.
Relevant at LinkedIn:
https://www.linkedin.com/posts/jurgenappelo_my-son-complained-this-morning-about-having-ugcPost-7327379208077275137-hjKc/?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAIrSTkBPM9XcceDcSGyfzZ_RouTiWLeC3g