Learning needs to be effortful to be effective(giansegato.com) |
Learning needs to be effortful to be effective(giansegato.com) |
I left a $20 bill in my dissertation at the library if you can find it. I did that to see if anyone would read it.
I dunno - I find it easier to remember when I type something out a couple of times. When using pencil and paper, I tend to forget since I am more focused on the writing bit than the remembering it.
Using copilot makes it much harder to pick up new programming languages cause you'll end up working on much higher abstractions, since copilot is generating a good chunk of the code.
Not to mention that your own code is much more sprinkled in, making it harder to build up a coherent, end-to-end understanding of the underlying syntax and semantics.
These superficial learnings essentially become my escape from real world: family responsibility, work and others. I guess I need to grow up and somehow indulge myself with real life but heck I refuse to do that even at 40+.
But until reading your comment I had never thought about how most people might just not have the energy to dig endlessly after every new idea. It didn't click.
You think you know when you learn, are more sure when you can write, even more when you can teach, but certain when you can program. -Alan Perlis
Whenever I want to learn something, after reading the textbooks and doing the problems I always try to implement something that uses that knowledge. For Math/Physics I can vouch for its effectiveness.
It becomes more relevant these days since we tend more and more to outsource content creation to the machines. This theory would predict that we would learn less doing so.
The worst part is the false sense of improvement. IMO, Learning nothing, yet feeling a sense of improvement is worse than learning nothing at all.
I know from myself that I tend to bookmark and save many interesting talks/videos and articles for later, but often I never end up revisiting them; information hoarding in some sense.
It works, it becomes a cycle of anxiety and blissful revelation, an obsession.
He clearly wasn't effective in his English classes.
>edutainment is not learning
I learned more things from entertainment than any textbook or classroom ever taught me.
Also, his choice of font is a blight upon typography.
effortful
adjective
showing or requiring effort
This statement is abstract enough that it can be both true and false depending on what one means when saying that. Can you be more specific? Have you watched movies and listening to music that gave you the skills to have the job or career you have? Did YouTube videos give you the knowledge to work in a scientific lab or start your own company?
Do you have any example?
There's a lot of educational content out there that's probably better to consume than dumb silly content, but skills and deep knowledge can only be acquired with practice, distraction-free studying, and similar activities.
On a tangential note, what does it mean to be “effective” in a class?
It sounds like you should have probably paid more attention to your books, and in the classroom.
If you can't get much value out of subject matter experts condensing the salient points of a topic into a book, with accompanying exercises that require you to master those points to solve...
It probably means you didn't do the exercises. Do the exercises.
Oregon Trail taught me more about American history than textbooks ever did, likewise Age of Empires II than textbooks ever did concerning medieval history. One of those Carmen Sandiego games did a better job teaching me English than any textbooks or classes ever did.
And god damn RPG Maker (read: computer programming) taught me more about math than any textbooks or classes ever did. To say nothing of all the games that demanded I learn how to do complex arithmetic in my head on the spot.
Looking back, the core lesson I took away is that you need to find the subject matter fun or meaningful to be able to learn it. No amount of "effort" is going to help if you simply don't, or worse can't, give a damn about the subject matter or the way it's presented.
How's your italian?
(「ペンネ」wa, oishii desu ne?)
Edit: for that matter, effortful has become more cromulent than I would've guessed: https://books.google.com/ngrams/graph?content=effortful&year...
While there are plenty of folks for whom this is true (myself included at times), it’s not a universal experience. I am good friends with folks who find joy in learning no matter how it is presented.
I feel there are two kinds of academic(or anything technical) knowledge in the world: One is of sugar, tastes sweet but does not provide much depth -- most youtube videos, blog posts and HN belong to this catalog. The other is of full course which provides in depth knowledge, and one needs not only to read but to try and practice based on them -- books, papers and some videos/blogs belong to this catalog.
This post is interesting, but it reads like a layperson (from a teaching point of view) coming up with some fairly good theories of learning (of which there are quite a few) and discussing how applying them makes concrete differences in knowledge acquisition.
If I was to make a guess, I'd say that the 'effortful' component the author discusses is most aligned with something commonly called the "transformation" stage of learning -- where the student takes things they have heard, seen, or experienced and attempts to transform some part of the world using them. A silly but apt example would be watching a youtube video on solving a rubik's cube, then having to transform the state of a real rubik's cube using your new knowledge.
This idea, and related ones, continue to be discussed at length by those who develop curriculum for educational institutions, and at a more meta-level, those who develop lessons for students who will later write a curriculum! I see a few links in sibling comments to Aviation learning, Youtube, PBS, but not to documents produced by any university on theory of learning. Perhaps they need to better link their work for smart hn readers to find.
Unfortunately, in my experience there are several shortcomings of the field (or at least typical bachelor's master's programs outside teachers's preparation, which is separate in Germany).
They love to talk about interdisciplinarity, and will even call out neuroscience and psychology often. In practice, it's mostly a social sciences field, and as such has been taken over by constructivism. Now, while learning is obviously (IMO) a constructive activity, they will "explain" everything with "handwave... social construction... handwave", while dismissing e.g. neuroscience aspects as "not a learning theory" and thus not worthy of discussion.
The statistical sophistication is very, very low. It's not all linear regression, but mostly, and even when it isn't there is no (even implicit) discussion of causality, just adding all variables (what McElreath calls "causal salad).
Fischer, K. W., Goswami, U., Geake, J., & Task Force on the Future of Educational Neuroscience. (2010). The future of educational neuroscience. Mind, Brain, and Education, 4(2), 68-80.
Mahadevan, J. (2020). Ethnographic studies in international human resource management: Types and usefulness. German Journal of Human Resource Management, 34(2), 228-251.
Romani, L., Barmeyer, C., Primecz, H., & Pilhofer, K. (2018). Cross-cultural management studies: state of the field in the four research paradigms. International Studies of Management & Organization, 48(3), 247-263.
Fortunately I see some very high quality and rigorous research techniques applied (with much less handwaving away of neuroscience) among my friends, but I don't need to look far for contrast. I don't work directly in this area, but am present for the pub discussions as it were.
However, I would say that statistical sophistication often seems to rely on the happenstance of cross-domain expertise. People only have so much time, knowledge, etc -- just the usual faults of humanity and organisations.
Some programs have really embraced neuroscience etc, but there’s minimal incentives for schools to keep up to date. Teachers going back for a masters don’t actually need anything more than a degree in hand to get a higher wage at their existing job. Which then feeds back into how these programs are setup and managed.
I guess you're studying "Erziehungswissenschaften"? You might find that educational psychology ("Pädagogische Psychologie") has a lot of the depth you would expect. One nice example is cognitive load theory.
One of the more complex mathematical treatments of learning are "knowledge and learning spaces". But it seems they get bogged down a lot in implementation details (= conflating implementation with interface behaviour). And: It's a bit to simplistic. For example, talking about things like the "expertise reversal effect" in this framework is not straightforward.
If I take a valuable nugget of information about it, is that real learning comes through experience: using the information, recalling it fron memory and manipulating it in some way. This goes from just writing what you hear to a personal project. Anything behavioral, more than just receiving and thinking, helps to actually learn.
I find it crazy that the educational system expects children to develop in this axis on their own almost entirely because they are simply expected to do so. I went to well-regarded schools and it’s generous to say they put a minimal amount of effort into teaching us how and why to learn. A frank assessment would be that they hardly seemed aware of the topic…
Sure, folks could just read the code, but inevitably people scan rather than read in detail. We all do that, and the cryptic nature of programming languages can just exacerbate it.
Typing naturally slows people down. It also provides a natural opportunity to make simple changes if they so desire.
Also, the article never mentioned the one absolutely knock down, finding that has been confirmed over and over again as effective for learning (or at least remembering): spaced repetition.
I believe you need to be conscious of what you are trying to learn and if it is tedious, then you are conscious of it because tedious means you experience some strain.
You will remember it better because it was painful. Just speculating but it makes sense to my layman brain.
Yes, people know. Education schools aren’t held in low esteem by social scientists or the educated public by reason of lack of familiarity. They can’t even manage a detectable effect on teacher effectiveness. Note I did not say large. The effect of an education degree on teacher effectiveness is not reliably distinguishable from zero.
> It's easier to pick a good teacher than to train one: Familiar and new results on the correlates of teacher effectiveness
> Neither holding a college major in education nor acquiring a master's degree is correlated with elementary and middle school teaching effectiveness, regardless of the university at which the degree was earned. Teachers generally do become more effective with a few years of teaching experience, but we also find evidence that teachers may become less effective with experience, particularly later in their careers. These and other findings with respect to the correlates of teacher effectiveness are obtained from estimations using value-added models that control for student characteristics as well as school and (where appropriate teacher) fixed effects in order to measure teacher effectiveness in reading and math for Florida students in fourth through eighth grades for eight school years, 2001–2002 through 2008–2009.
https://www.sciencedirect.com/science/article/abs/pii/S02727...
They deal with very specific constraints because of classroom economy.
If your stated goal is to get every child in this room to read, then you're wildly out of alignment with the needs of the child who could be reading the classics by the end of the year.
Gifted and accelerated programs only alleviate the problem by a fixed amount by adding one additional more intense tier, which every child competes for to improve their university application profile.
Any person who tried to teach a child one to one for a short period of time learns how unfit school is. The only limit to their learning is the adult's energy and skill. While school is the product of several rounds of the lowest slowest denominator.
Kings and princes used to be taught by tutors they form long term relationships with. This is still the best way. School is the equivalent of feeding the masses dog food to keep them compliant.
One of the earlier posts decried "constructivism" in research, but I don't know how you get around the fact that most students are not, even cannot even be allowed to reach their full potential in whatever way that applies to them, because society requires us to fill certain roles, which that potential may not apply to. Educational outcomes are socially-constructed; I'd even go so far as to say that this social need is more predictive than inborn potential and even effort.
I was probably one of the handful people who were actually paid to program Apple IIs in Brazil.
It was a remarkable experience to work at a applied research centre early in my career (I was in my second year in college) that had a lasting impact on everyone involved.
It also argues that easier/faster learning is often worse. Harder learning lasts longer, and a big part of it is what they refer to as finding ways to "interrupt the natural process of forgetting". With this in mind, interleaved/ varied practice, spaced repetition, reflection, are all techniques that both lead to deeper understanding and better retention.
There's analysis done on the "teacher quality" (based on student results) and the amount of teacher training (at these schools). There's basically zero difference between an emergency qualified teacher with no education qualifications, and someone with a masters or PhD of education.
Quote - https://caldercenter.org/sites/default/files/1001059_Teacher...
> Our results suggest that only two of the forms of teacher training we study influence productivity. First, content-focused teacher professional development is positively associated with productivity in middle and high school math. Second, more experienced teachers appear more effective in teaching elementary math and reading and middle school math. There is no evidence that either pre-service (undergraduate) training or the scholastic aptitude of teachers influences their ability to increase student achievement.
The evidence is that teachers learn literally nothing (or overall nothing - maybe they learn some good things and some harmful myths) in teacher college. They might learn a bit of math, which helps them teach math. IIRC it's also found that doing a PhD in education can be helpful in teaching classes with a heavy essay writing focus.
It's not surprising, since a lot of what gets taught in teacher college is junk psychology from the 1930s, and the essays of the kind of people who do a PhD in education. Here's a few of the biases that they tend to have (IMO):
* Education experts did OK at school, and think that the fundamentals are boring because their white upper-middle-class parents helped them learn phonics and the times tables.
* Education experts are often not actually good at much other than essay writing and year 10 math, and don't see the point of hard sciences like pyschology.
* There's plenty of smart people who go into education, but they do it because they want to teach, not be a researcher.
I could go through some of the poor content in education degrees (like the hero of education experts Vygotsky, who I'm not convinced many have even read other than through n-th hand sources where people write "the student is thus about to construct knowledge through scaffolded tasks in the zone of proximal development (Vygoskly, 1930something)", but you'd be lucky to find many teachers who agree on what any of that means (other than be common sense "don't make it too hard to follow").
Is it really a wonder that Cletus, who got a job in rural school in Alabama with no more qualifications than being able to write his name correctly teaches just as well (according to research) as someone with a PhD in education?
The good research is in psychology departments - people who study the science of learning are generally pretty well informed on it.
That's probably why some school subjects that lack this component largely or entirely like history, geography or literature are so hard to learn.
Hear and forget, see and remember, do and understand.
Is this a reference to some specific university?
Learning Is an Active Process Learners do not soak up knowledge like a sponge absorbs water. The instructor cannot assume that learners remember something just because they were in the classroom, shop, or aircraft when the instructor presented the material. Neither can the instructor assume the learners can apply what they know because they can quote the correct answer verbatim. For effective knowledge transfer, learners need to react and respond, perhaps outwardly, perhaps only inwardly, emotionally, or intellectually.
[0] https://www.faa.gov/regulations_policies/handbooks_manuals/a...
Nowadays, I need comprehensible input the whole time, hooks to actually understand what's going on. I'm not sure if it's an issue with brain plasticity or growing impatience, perhaps a learned discomfort of not understanding things. Perhaps the internet has since trained me to google things instead of stopping to think or just soldier through, learned helplessness resorting to the hivemind's omniscience if you will.
I suspect this refusal to engage with the material and just flee to the search is this very lack of effort, which has caused me to stagnate. Alas...
After a bookish childhood, I spent years as an adult trying to "learn more." Mnenomic tricks, Anki, active note-taking, exercises, trying a different book if the first one doesn't stick. It didn't work very well and I probably wasted a lot of time. Two things became clear: 1) I was focusing on the process as a way to procrastinate on the real effort involved, 2) I was compensating for a lack of real curiosity about the subject matter. Yes, learning is hard work, but does it need to _feel hard_? When I've felt truly engrossed in something, I don't need to remind myself to exert effort, and I'm not really thinking about whatever X% of learning gain I can get from doing it better.
Deep down, I think like seeing myself as the kind of person who is "really into learning" -- math, theology, classics of literature, CS, art history, whatever. Keeping up the identity I developed as a nerdy kid. Of course, it's important to be gritty about learning because... why? Hustle culture?
Now, I'm trying the opposite approach: enjoy as much entertainment as I want, avoid exerting any effort or discipline in learning, and stop immediately if I'm not feeling it. Part of this is not beating myself up when I naturally lose interest in something (95% of things). Yes, it's easy to get distracted by low-effort scrolling and such. Ultimately, though, avoiding exertion makes it easier to focus on those rare things that truly spark wonder.
Conscious effortful study can help you pick up new vocabulary and grammar, but our brains have to process language at speeds too fast for the conscious mind to keep up. It has to be learned subconsciously. Most of us don't even notice the effort needed to read this paragraph of text because it's all subconsciously processed before out conscious minds get to it. While conscious learning requires short bursts of focused attention, subconscious learning is done almost 24/7 in the background in a relaxed state (no effort is needed, and I'm not even sure if it makes a difference).
[1] None of this is backed up, I essentially pulled it out my ass. It's my theory on why language classes don't work, but children pick up languages so effortlessly.
It is one thing I hope AIs can really help with. It is a common belief that you don't understand something until you can explain it to someone else in simple terms. I've tried this method with AI chat bots with some success. They have two immense advantages over other humans: infinite patience and bias towards understanding over agreement.
For complex and abstract concepts, most people just don't want to sit and listen to a 40+ year old dude mansplain them. AIs - no problem. They'll listen forever and never get tired, bored, frustrated, etc.
The second advantage is even more important. Most people listen only to wait for their turn to speak. Or they will get caught up on a minor point because they don't agree. That proclivity to disagree can often turn into blindness, erasing any further information past wherever they got hung up. Even if they do ask you questions, often it is some attempt to highlight their disagreements, in a pseudo-socratic method kind of lawyering. AIs have no ego, they don't have stake in the game. They aren't trying to convince you to agree with them. They can just listen and understand, rephrasing and repeating back.
https://mathshistory.st-andrews.ac.uk/Biographies/Halmos/quo....
This core idea is getting at something important, something that other commenters are covering and is covered in this previous HN thread [0] about information addiction for example, but I disagree the author's assertion that all passive consumption ought to be categorized as entertainment and all active creation ought to be categorized as effortful work. (I've seen too many video game damage calculation spreadsheets for that to be the case.)
I'll highlight the author's conjecture that "edutainment is not learning but preparation for learning". Relevant, digestible, and yes sometimes entertaining collections of information are preparation for understanding (which typically requires application), which is preparation for mastery (which typically requires ten more years of application). I would argue this entire process is what encompasses learning, of which well-sourced information is a critical component for. I suspect there is some conflation here of entertainment with the risk for distraction, which is a real mind killer that ought to be addressed, but instead gets tossed away by the author during his Cal Newport reference in favor of his love for Twitter and vested interest in newsletter subscriptions.
The concept I feel the author is getting at via his edutainment strawman is that information acquisition is not sufficient for the fluency of understanding required for conceptual mastery. This is a concept that I think most HN readers and textbook exercise writers would agree with.
The possibility that the author may be missing a working understanding of this concept feels to me like it would explain a certain awkwardness about the entire article, which seems to rely on shoehorning a plethora of loosely-connected, name-dropped quotes and ideas into italicized slots of questionable logical integrity to support the presumption that everything entertaining must be useless, and everything educational must be hard. I've met way too many lazy smart people to believe that to be the case.
E.g. at high school I learnt mathematics purely from reading textbooks, and found the requirement of needing to do homework and practice exercises in class hugely frustrating. For context, I did well and went on to do mathematics at university level, where my learning style more or less stayed the same.
While a great friend (who is quite possibly smarter than me) is the opposite, and they were greatly frustrated by the lack of putting things into practice, and felt it hindered their learning.
At the end of each chapter, there's usually a mostly-blank page. That's your spot to summarize or react to what you just absorbed.
The only downside is it requires your own dead tree copy, and anyone who borrows it later will inherit your goofy notes.
I've even tried printing out blog articles that I want to read more actively (2 pages per sheet, double sided).
I ended up buying a Supernote A5X and it has proven to be one of the best purchases I’ve ever made. The Digest feature is fantastic for the workflow you’re describing. Ripping the DRM from Kindle books is trivial and you can export them with Calibre to include oversized margins if you’d like more space.
Writing in a book is abhorrent to me, but it's a personal preference.
The biggest advantage I see in this approach is that it works with multiple mediums and provides sufficient space, flexibility of formatting and ability to search.
Of course, feeling of a real book in hand is something else.
... I also think that this is usually a difficult experience to guide people into, especially in a classroom. In many ways, the classroom seems almost designed to prevent this state:
* You're surrounded by people, often noisy
* You do not often get to choose your subject matter
* If you zone out (ie, zone in), and do something weird (sit funny, pick your nose, rock back and forth, doodle, whatever) then students and even the teacher will break your flow. This means you need to be watching yourself, which is antithetical to flow.
* Regular interruptions, lack of physical activity, disconnection from nature, etc, etc.
* As the author very rightly points out, the lack of movement in most 'learning' situations is more harmful than most realize.
This leads to people associating learning with work, with effort, with strife. People decide they don't like math, or music, or art, or reading, because their school experiences are so, so bad.
It's kind of grotesque.
Our app was originally built with students in mind, and when building it I wanted as much as possible to build a tool that helped students learn. Not just make them feel like it was helping, but actually helping. Although we are no longer specifically targeting students, that goal is still in my mind.
So we have been slow to integrate "AI" into our app because I worry about the effects of having an LLM do an appreciable amount of your writing for you. To me, an important aspect of using a tool for knowledge management is to actually do much of that management yourself, otherwise it won't stick.
There are certainly a number of ways that I think LLMs can be quite useful without hindering that goal, but I need to spend a bit more time working that out for myself rather than just jumping on the "let GPT-4 finish the note for you" bandwagon.
I think boredom played a huge part in my learning. 1970s and 1980s at home there wasn't much to do. Any new material or idea was like water in a desert. We didn't know it at the time it was just normal.
Even early days of home computers it was all you on your own learning how to fix anything. Even early days of Internet it was just on for a bit and log off even shut down the computer and cover it with a cover. Now it's 24/7/365 stimulation home or outside walking from a fire hose of knowledge.
Things you learn;
- Small text is hard to read
- Making things hard to read means they get missed sometimes
- Making it the only thing on the screen makes people try to read it really hard
- Which makes them more likely to remember it later
I work in UI/UX design and this f*cked with my brain for a while. It's counter to everything we're being taught.
Human brains are fun.
Interesting to hear this. For me it's the opposite; while listening I conduct mini arguments with the presenters in my head, and come up with counter arguments, and counter-counter arguments, and counter-counter-counter arguments.
But a better habit is first check if it matters. Most of the time standing back and asking "why are we even talking about this?" - is the correct question, because usually there's no good reason; it's 'news' aka noise (definitely not education).
The concept of a "learning box" sounds interesting, and it would be great if there were an extension for it.
On learning, what I have learnt without effort, but with the lesson of time and painless observation: it is harder to unlearn.
And since content is virtually infinite - far more content was produced today than anyone could consume in 20 lives - it is crutial to currate. Given the amount of noise out there? Sharpen great and fast information filters.
If an article makes a statement with little to no clear argument to back it up, I continue to read but faster, parsing only one every two or three words until the end of a section.
If the same statement is made again without any backing, that's potential fallacies to brainwashing. An author dragging its feet for ads revenue or filling its prints doesn't fall into the conscious brainwasher category but still represent the risk to spoil my time at best.
If no decent argument is laid out by the end of the section, to back the initial claim, a beginning of one at least, then I stop reading the piece. There statistically isn't enough value in there to burn time reading. Or worse, fallacies may start to do their magic.
I did take the time to skim through the entirely of this article, just to confirm the conclusion made after reaching the first part, to write this comment
- Make some kind of logical scheme to help with disambiguation (things like which legal process comes first, of the Socrates-Plato-Aristotle order of philosophers and ideas)
- Write the long form of the subject everything by hand. Pen on paper. If it's a law or bill, write it fully, if it's a scientific concept, write every part of it that is related to the subjects that can appear on the test.
It's slow, tiresome and a complete nightmare for students with anxiety problems (trust me, i know), but as you write, you very slowly burns the knowledge that fills your logical scheme into your brain.
That's what the people that succeed in hard tests do to dominate nebulous topics. I don't know if it translates to actual mastery in different areas, but it's a very strong studying technique
Much like the actual blog post.
In fact, the only things that I have poor retention of is Algebraic Geometry because despite the amount of time and effort I spent on it I made no progress in intuition. Partly because I was attempting the "take notes" strategy for that class instead of my usual "high attention / no notes" strategy.
Learning is an investment, and as my salary increases and life gets more comfortable, my brain doesn't feel the urge to learn and prefers leisure.
I'm looking for ways to "hack" my brain and get back into the habit of learning in my free time and wanting to figure things out. When I have kids I hope they will be curious and I'll learn with them.
Please share anything you've found. For me:
- discovering Obsidian a year ago was a great boon. It led me to comb through millions of words of old person notes, remembering many different me's interest in different topics etc. When approaching a new topic, I now compulsively take detailed notes and then reform them into "evergreen notes" (google this term) although it's not the same effortless joy of my childhood.
- recently refinding old sites like wiki.c2.com has rekindled a great passion. The style of discussions, without marketing, status seeking, linking to blogs, medium articles etc. is also extremely pleasant and lets the content float up much better. Even modern communities which don't suffer from these problems still seem to have jaded userbases, who are just... Tired and not willing to really revel in their knowledge, but protectedly proactively avoid things.
- really calming myself down before engaging, e.g. closing my eyes for 10 mins in quiet, or just doodling on a paper, maybe stacking some 9 volt batteries into a tower and calmly approaching the topic, reading or such. In this way, reading can almost be like a reverse stream of consciousness. However life quickly knocks, taking me out of this (flow?) state. I recall some researchers discussing how children effortlessly play (so the opposite of the original topic about effortful learning...) experimenting, collecting data etc. to learn and develop their worldviews, but if you impose requirements or expectations on them, they learn slower and don't blossom.
I also make "grab bags" for everything, I am a perpetually disorganized person but this lets me be prepared. I have a drone bag, an electronics box, a bike toolbox, any significant project gets a bag/box I grab on a whim. That way I lower the preparedness bar and boring organization tasks.
I have _NO CLUE_ about 80% of the topics on the more complex subjects. Especially when he's explaining the equations. I rewatch the videos sometimes 4-5 times and still learn things that my mind glossed over. Lots of "ah ha!" moments. Love this channel.
I found the experience enlightening.
It may well be the case that 'comprehensible input' is enough for languages because we already _understand_ at least one language to a functional degree, and there's no real 'learning' a foreign language, moreso _remembering_ words and structures. Much like most people don't 'understand' the mechanics of their native language, they merely use them.
I don't think that we can readily assume that comprehensible input is thus generalisable across other things we would wish to learn.
- The amount of comprehensible input needed is really huge. So much so that someone who has 10-20 hours a week to learn a language might be better using an active approach.
- I used to think that if you understood a language certainly you would also be able to speak it. But this just isn't true. Passive vocabulary and active vocabulary (not to mention sentence construction) are separate and need to be practiced individually.
- In my own experience I have been very successfull with comprehensible input when paired with being in the country or studying the language during the day. This allows me to solidify links from what I've heard and read with the real world. AFAIK, his studies had a similar setup, always including students who were studying the language and then did extra reading/watching on top of that.
There is a bit of a pipe dream, that you will just watch mangas all day at your level and suddenly come out fluent.
I wonder if the nature of internet discourse has also changed, reducing how much there is. Comparing wiki.c2.com style pages/discussions with reddit now, things evolve in a much less coherent manner, there is less of a conversational back and forth truly developing ideas and more of a show-offy approach. Perhaps topics are also more difficult - e.g. learning CSS today with 30+ units of measurement must be hell compared to back in the day.
The safer bet. Children and adult cognitive loads are assigned different in Maslow's hierarchy of needs.
> μὴ εἶναι βασιλικὴν ἀτραπὸν ἐπί γεωμετρίαν — E
* of course, a major part of this learning process involves doing the exercises, and if no one has been nice enough to set any exercises, then one simply has to provide one's own by playing around with other combinations of splodges to see if anything interesting results. cf http://www.willamette.edu/~fruehr/haskell/evolution.html
One possibility is that it becomes more challenging to integrate information the more you know, and the more that you are aware of your limitations.
The child might read something and pick up the general idea of what things are from the context. An adult might realize that the context definition is incomplete.
A child might read an instruction that says to use a box wrench to tighten the bolt and pick up from contacts that that's just the tool that you use.
An adult night Wonder what kind of box wrench, metric or imperial ratcheting or not. What about other times where I use the socket for this type of work, is that better
Look at any activity, chess, music, physical sports, at top levels most of it is subconscious effort.
Imagine having to constantly look at your keyboard, consciously searching for the letters to communicate your thoughts.
How would you expect to communicate well if your conscious process includes the menial search for the individual letters? How would you be a good chess player if you didn't have an intuitive feeling for most board positions but had to evaluate every move for a long time?
The things you learned well, were things that you successfully committed to your subconsciousness. Sometimes you plateau and think there's no room to improve, but by inspecting the conscious effort, you can find things that need to be practiced to make them subconscious.
However, that doesn't mean that the process of acquiring the language isn't effortful! Even the most hardcore no-interface positionists accept that the subconscious learning system uses linguistic input that is _processed_ for its meaning and _comprehended_ (called "intake" by some experts) as its source material. When you don't have automatized knowledge yet, comprehending is very effortful and cannot happen without focused processing. The theoretical debate is more about whether that processing requires "metacognitive" processes and declarative memory and automatizing those, or is just getting intake (input comprehended by one means or others, without regarding any "rules" or "deduction" or "explicit/system 2 thinking") enough. But all in all, it takes effort anyway, in the sense that you still need to focus, and you might get frustrated of not understanding enough, struggling to follow.
Also, there are some studies that claim that children don't pick up languages as effortlessly than is generally thought. Quickly? Yes. Often with good results (in terms of pronunciation and grammar)? Yes. But effortlessly? Not so much!
Children may appear to be picking up languages effortlessly, but consider how much time they spend to do so: take 2 years of a language, and you will be in advance of a 2 year-old native speaker.
> our brains have to process language at speeds too fast for the conscious mind to keep up
good point for oral communication, but literacy doesn't have to be done in real time
They’ll try to coerce any of your (or the model’s) mistakes into the correct answer out of politeness, and when you make a mistake it has zero model of the model you used to produce such a mistake, so it has no idea where to intervene in your knowledge except at which words hit the keyboard.
Idk, I am extremely extremely unimpressed any time I try to quiz it on topics I know. Except for coding, which makes sense since that’s the exercise of pushing as much semantic information into syntactic information as possible.
I try to explain a subject to the AI as if I am trying to teach a friend who has no knowledge on the subject. Then I am judging how well I can communicate the idea. In many cases as soon as I try to explain the idea I realize that my own knowledge isn't deep. So I am judging myself, not the quality of the AI.
An analogy is the Rubber Duck method of program debugging. Most programers have been in the situation where when they try to explain an answer to another programmer the answer will suddenly pop into their head. The value isn't the knowledge in the other programmers head, it is in the act of trying to explain. LLMs are patient and unopinionated and make for a good recipient.
My friend seems to operate with a strong mental scaffolding that allows him to quickly absorb new information -- you can almost see his brain saying "this goes here, that goes there" as you speak to him. This scaffolding also plainly exposes holes in knowledge, which makes him an incredible conversation partner when you're talking about a technical topic -- "you mentioned A, B, and D. Seems like there should also be a C. What happened to that?"
I've known this guy for a long time, and you can imagine that his scaffolding only grows stronger over time, so that it's very rare to discuss a topic with him where he's completely adrift. It helps that he doesn't have the personality of a know-it-all; even if he is a master of a topic, he has an uncanny ability to uncover the thing you know that he doesn't, which makes conversation with him enjoyable rather than intimidating.
Two final points.
One, his memory is not notably better than anyone else's. He forgets conversations and situations as much as anyone else, but he definitely remembers when he's filled a hole in his scaffolding.
Two, his mental scaffolding is very conventional. By that, I mean he rarely has a weird way of thinking about things that makes sense to him but to nobody else. So when he asks a question, it's almost always the same deep question that other experts in a field have been working on, rather than a strange take. While this seems like a liability, it means that his conversations with experts tend to get to the important parts very quickly, rather than wasting half an hour establishing a common lexicon. Maybe I'm confusing cause and effect, and the conventionality of his scaffolding is a consequence of the vast amount of knowledge he's learned and needed to reconcile, rather than the reverse.
And interest.
I've had several doctoral students and teaching assistants tell me (with variations) that they aren't good in "that math stuff" and that they don't think it really matters; they are doing "best practices" in SPSS and it's much more important to look at it in a holistic way.
In theory they would be right. If the best practices were sane.
But I also think it is going to get better, only very slowly. At some point there has to be cross-pollination with other fields.
I suspect that even a tin-eared adult learner could master those phonemes with enough obsessive repetition, both listening and speaking. I'm reminded of Liam Neeson's character in 'Taken' spending a transatlantic journey listening over and over again to the voice of his daughter's kidnapper just so he could recognise him later on.
That level of obsession requires motivation which I think is highly under-appreciated where learning is concerned.
We often think of small children as possessing magic learning brains without considering how highly motivated they must be. The ability to speak means they can more easily make their basic preferences and problems known to parents and so on. It's hard to imagine what it was like for us back in early childhood.
However, children do have one natural advantage: if you learnt a language when young then it must be harder to forget it, due to Ribot's Law:
no effort: your L1, their L2
easy mode: your L1, their L1
aim for: your L2, their L1
hard mode: your L2, their L2
> children are very motivated
That, and much less afraid that they "might make a mistake"
A two-year-old lives in a full immersion environment where the only language they can communicate in is the one they are learning and where a large portion of the people they interact with are going out of their way to help them practice. They also have no responsibilities other than learning language and a few other skills.
Put an adult in an environment like that and they will make extremely rapid progress.
You're correct that it's an exercise in introspection, rather than relying on the AI's own knowledge. It was clear to me when trying to write explanations to its questions where my understanding was starting to get fuzzy and hand wavy.
A nice bonus was to get it produce a scorecard of correct, nearly correct and incorrect explanations. I could see these as a good jumping off point for me to do more learning/research. Though I suspect the AI would be less accurate at this in a more niche topic than I chose (refrigeration).
I've been asking the chatbot to point out my assumptions and to point out weaknesses or errors in my reasoning. That has been pretty useful in uncovering aspects of my thought that aren't as foundational or coherent/consistent as I expected. It also doesn't require the AI to be an expert in order to give me feedback, it just points out unchallenged assertions and errors in logic.
I think this is a potentially powerful use for AIs. They can patiently and politely nitpick and point out subtle errors in our thinking process, providing unbiased critical feedback. That is almost as valuable as having them be perfect oracles that can answer any possible question.
Thank you for sharing. It helps that refrigeration has always sort of blown my mind, so this was an interesting topic as well.
Same! It's been in the front of mind because I've recently been watching Hyperspace Pirate's videos on building a DIY cryocooler (and have been struggling to follow it at a technical level). You might enjoy them too: https://www.youtube.com/watch?v=7QZrHzd3RA8
If anything, I'd argue the pain and effort can be negative it will still work.
This is not true. Anki is tedious and causes pain. Flashcards are such in general.
But dpaced repetition does not requires either.
Flashcard are just easy to implement application of it.
Yes among flashcards, anki is the most popular. But naive use of it leads to a lot of pain and tons of problems, you have to learn how to use anki.
Actual memory "devices" (tricks, schemes, etc.) are orthogonal; SRS (for me) just optimizes my time so I don't have to go over shit I know more often than I need to.
For anyone curious, my 2 go-to books on actually learning how to remember are the ones by Higbee (dry, but thorough) and Lorayne (approachable, but feels less academic). I used SRS to remember those 2 authors w/o having to look them up, FWIW. ;-)
Default settings are ok although I'd say 2 things that are counter to some common advice there:
* Actually use all 4 ratings buttons, and come up with a scheme for when. For me it was 1: missed it, 2: had to guess, but guessed right, 3: got it, 4: got it without even reading the entire prompt
* If you're insistent on the (IMO bad advice) "only use 1 and 3 keys" thing, install the "Straight Rewards" plugin. Without it, all your ease factors will decay to their lowest level (130% by default) since '1' lowers it, but '3' does not raise it, so it can only go down. With "Straight Rewards", it'll up the ease factor for you gently as you rate a card '3' enough times in a row.
Alternately, install the FSRS4Anki plugin. It's more work, but is a more dynamic scheduler. Requires the V3 Anki scheduler being enabled, and one of the more later versions. Do some reading on this before you try it; you can always go back, but it's not been proven yet to be substantially better than SM2.
That's not really SRS though; a key part of SRS is the heuristic for when you next get exposed to the thing you're trying to remember based on your previous "got it"/"missed it" response. I suppose "every day" is one such heuristic, but it's not really in the spirit of SRS.
And you're mixing up enjoying the content with the mechanics of an SRS system.
If you really WANT to learn something you cannot rely on such accidents happening. Accidents are not a method.