Oh god, the quoted story from "Spent" was a total win. =D

That and the whole "civilization" and "progress" thing made me think of David Brin's short story "NatuLife(R)".

I know you don't like scifi (or camelCase xD), but you should read it and compare and contrast!

Page 147:  http://libgen.info/view.php?id...

(Also, yay, the straight-to-my-gmail update notification thinger I set up at http://blogtrottr.com/ for your 'daily' log worked! ^^ )

Are you referring here only to psychological behaviourism or logical behavriourism (theory of mind) as well?  I find the former extremely plausible (having sat a fair bit of vipassana, it starts to look just obvious, in fact), but the latter not quite as compelling -  if you do find it so I would be interested to know why.

I'd generally extend it all the way up to logical behaviorism, but the more universal the claim, the less certain I am of it. :)

Overall though, every time I thought I had "something left over" that a behavioristic account couldn't cover, I turned out to be very wrong, so even though I still have a few hold-outs, I've changed what horse to bet on for now.

(That doesn't make it a *full* theory of mind (or ontology) yet because there are still some crucial black boxes left, like what exactly a "cause" is, or how an agent detects "sameness". But these problems are shared by all theories so far.)

The primary thing that did convince me is just actually reducing and engineering learning and instruction. After internalizing that "understanding" isn't magic, but follows logical, functional laws like everything else I know about, made stuff like "dispositions" and "intentionality" seem really weird. The cognitivists (or worse, mysterians, dualists and whatnot) just don't have anything interesting to offer. I have a high tolerance for (temporarily) unreduced magic and odd naming conventions (like tantra), but it better actually *do something*.

It's an insane life cult, not an insane death cult!

More seriously, I think the best argument against anarchoprimitivism - beyond the fact that you just can't support 7 billion people on it - is that it's a merely local optimum and to get to a global one we'll have to push through the valley of the shadow of death. But maybe that's just my Marxism talking. 

(Also, apropos of nothing, started Feser's Aquinas. It's well-written as expected, but I fear to actually understand the stuff I'll have to wade through Oderberg and ugh.)

@a695e7582563edc288c4fd3587556624:disqus 

Quoting the conclusion from a chat I had about this topic:

Me: Absent any strong evidence for [the impeding Eschaton], I'm going to side with Malthus and the 2nd Law of Thermodynamics. :)
Me: And the whole thing is a bit of a distraction anyway.
Them: From?
Me: Because "civilized life sucks" ignores that civilization doesn't just compete with primitivism, but it does entirely new things.
Me: So yeah, all the things that both do, hunter-gatherers might do better. But they're still barbarians, so fuck them.
Me: Noting that hunter-gatherers enjoy their life more than farmers, and wanting to fix that, is like noting that giraffes have longer necks, and wanting to genetically engineer new humans to fix this discrepancy.
Me: Which misses the point that you're competing with stupid long-horses.
Me: Meaning, civilization does *new* stuff, which comes at a cost. You might be able to lower it, but if you're *primarily* focused on getting the cost down, compared to someone you ought not compete with, you'll screw up the good stuff.
Them: Obviously, we're not competing with hunter-gatherers to enjoy our lives more, so...?
Them: I'm not sure I get the giraffe analogy.
Me: Wasn't that the initial argument?
Them: Huh?
Me: That hunter-gatherers enjoy their life so much more, and we should try to have at least as much fun?
Them: Oh, you mean that's exactly the point you were making?
Them: So confused now. xD
Me: I'm saying that a) it seems unlikely we will enjoy our lives more than hunter-gatherers (except maybe temporary), and b) we shouldn't try to.
Me: The giraffes were an example of some other animal doing a better job at some criterion that's irrelephant.

(If you make any sense of essentialism that doesn't require a denial of biology or "reality is a lie" harmonizations, tell me, btw.)

You don't have to strip my name from anything, dude. xD (Well, except for, you know, that one time with the ducks and stuff, when we were plotting to assassinate the president or something...)

I'm still not sure I get the point you were making.

Yes, it is irrelevant whether or not hunter-gatherers did a better job enjoying their lives.

Because even if, hippothetically, being a hunter-gathered had sucked, I would still have the same goal to make the world *fun*.

Hunter-gathers just provide a baseline to start by beating, and a little more information about ourselves...

I've read "Don't Sleep There Are Snakes", and sure, I believe the Piraha are *happier*... but I wouldn't want to *be* one; They're kinda jerks. xD

@Owen_Richardson:disqus

(Just anonymized out of habit. Noted.)

I'm not sure if we're just violently agreeing or not. I'll try to re-state my full point. (I'm also not attributing any of the potential criticisms here to you.)

The anarcho-primitivist critique I originally referenced is a reaction to stuff like this:

(distilled caricature mode ON) "Civilization is awesome! Medicine! Not living in caves! Look how amazing science and technology are!" (distilled caricature mode OFF)

To which it replies: "Well... pretty much none of the factors that we'd normally associate with life satisfaction are actually *better* in post-agricultural society for the average person. In fact, many have only very recently become about as good as they originally were. And civilization *adds* a lot of problems."

There are multiple ways to respond to that. (The list is neither exhaustive nor exclusive, of course.)

1) "Yes, but we now have many more people, and life is still good *enough*."
2) "True, but there'll be an Eschaton eventually that makes up for the current suckiness."
3) "We should abandon civilization to the best of our ability."

People with a stronger antinatalist bend would disagree with 1), but I'm mostly agnostic about it. I don't have enough data to decide that either way. I'm also not particularly convinced by the implicit utilitarian argument.

I'm sympathetic to 2), but I'm fairly skeptical of it. It seems to me that the reasons that make civilization suck are economic and game-theoretic in nature. Changing incentives is really damn hard, and no technology that I currently consider fairly plausible could deliver on that.

Someone who thinks AI / behaviorism / Marxism / prediction markets / praying for Jesus / a God-Emperor etc. could actually overcome Malthus will of course disagree, and I don't want to ridicule or dismiss this option, but I'm by-default skeptical of any proposal that requires large rule changes, in any discipline.

Lastly, I think 3) is completely infeasible and won't happen ever, unless some really precise apocalypse hits us that is just strong enough to make farming impossible once again, but not strong enough to kill us all. Which is less likely than the God-Emperor option. You might as well try to convince people over the internet to stop having children.

However, my main point of the giraffe analogy was that 2) and 3), to some degree, miss the point. The life satisfaction of foragers is simply not relevant for us, in the same way that the neck length of giraffes is.

The unique advantage of civilization is that it generates *meaning*. It produces things like honor, St. Peter's Basilica, and Godzilla: Final Wars.

If you engage in the primitivist / anti-primitivist debate, you're implicitly granting that life satisfaction (in this sense) *is* a relevant variable. It isn't.

Imagine there was a debate between pro-giraffists and anti-giraffists. You reply, "guys, I don't think neck length is normative, so could we stop comparing ourselves to giraffes?", for which you get accused of wanting to live without any neck at all, you anti-neckist. (Analogous to being accused of being pro-suffering for rejecting "happiness" as normative.)

The correct response to "based on hunter-gatherer standards, our lives suck" is "true, but who gives a shit?". They aren't even informative as a baseline because they aren't doing the same thing.

So in a sense, I'm siding with the civilizationists, but I think their "life doesn't suck" / "life won't suck soon" argument is crap, and they shouldn't use it. That's like replying to "my game is better because it made more money" with "well, my niche isn't *that* small...". No. We're not in it for the money, or the reproductive advantage, or the comfort. That doesn't mean those things are unimportant, instrumentally speaking, but they aren't the *goal*.

Civilizationists otter know better.

(This is also why I don't like the use of "happiness" anymore. "Being happy" is a useful behavior, but not the thing itself you aim for. Treating "happiness" as the goal is like treating "the clicker" as the goal of clicker training.)

Hope that clarifies my semi-rant.

Glad you are back blogging—I've missed you!

The one thing that I found surprising and interesting in Spent—although it borders on crankish—is his theory of memetic vulnerability. The idea is that cultural conservatives accurately diagnose themselves as vulnerable to infection with stupid ideas, and so reject everything new. People who are memetically vulnerable and fail to recognize this wind up as New Age airheads or worse. Apart from that, the book is a good popular introduction to signalling theory.

Yeah, over on LW, they call this Reason as Memetic Immune Disorder.

I'm just hesitant to actually take any argument based on memes seriously until there's even a hint of a mechanism for how memes could actually work. :) (Using them as a useful fiction or shorthand for "ideas with selection-like characteristics, sure, that's fine.)

It seems to me that your attitude towards "beliefs" depends primarily on what interpretation for the belief game you've learned (or are willing to use in a given context) - belief as attire, belief as prediction, etc., and that most of those "pathologies" are simply the result of a mistaken interpretation.

In the Theory of Instruction literature, they give the case of a bunch of otherwise smart highschool students that never learned that given instructions are actually *relevant* for the problems they face and not just a useless ritual. Quote from Teaching Needy Kids in our Backward System":

A related theme that Greta and I stressed was that there is no such thing as a general understanding. If you don't understand the details, you don't have a clear understanding of the whole. Most of the students were poor in math. Part of their problem was that they were not practiced in attending to the details presented in directions. During the first few days of the summer program, the students showed all the signs of attending; however, their latter behavior revealed that they didn't pay attention to the directions. I presented. Instead, they did what they assumed they were supposed to do. If I told them, "Copy the equation for the next problem, then stop", many would write the equation and continue working the problem.

For most of the students, written directions were not a specific guide about exactly what to do, simply rituals that accompanied school work. The typical approach students had was to ignore the directions and try to work the first problem the way they thought it should be worked. If they succeeded, they would work the rest of the problems in the set the same way. If this strategy failed, students would raise their hand. Their teachers would typically not refer to the directions but would show them how to work the problem. This demonstration served as a model for how to work the rest of the problems in the set.

The students' behavior implied that they needed direction instruction in following directions and that they needed a lot of practice. The problem I created took the form of arbitrary and low-probability directions. The first example was a sheet with a set of simple two-digit addition-subtraction problems, and the following instructions: Work the problems in this order: First work problem 5, then work 7, then work 3, then start with 1 and work the rest of the problems in order.

I told the students, "Everybody, touch the directions at the top of the page. Read the directions yourself. Read them carefully. Raise your hand when you've read them and understand them."

Hands went up. "Listen: Read the directions one more time. Raise your hand when you've read them and understand them."

Hands rose. I said, "Work the problems the way the directions tell you to work them."

The responses of the students showed how strongly ingrained their misconceptions were. Most of them worked problem 1 first, the n2. I presented questions to students who didn't follow the directions. Very close variations of the following routine occurred with at least eight students.

"Stop. Did you do what the directions told you to?"

"What they tell me to do?"

"They're at the top of the page. Read them out loud."

After the students read the directions, I said, "Did you do that?"

"Do what?"

"Do what the directions told you to do."

"What do they tell me to do?"

"Read them out loud." The student read them again, and again I asked,
"Did you do what the directions told you to do?"

"What do they tell me to do?"

"Read them out loud." ...

Usually, after the third round, the student would look at me and say in disbelief, "You mean, you want me to do problem 5 first?"

"Yes".

Although most of the students had the same confusion, not one student in the class questioned the meaning of the directions before working the problems.

We continued to present arbitrary instructions for working problems until the students became quite reliable in following them.

It goes on for a bit more, and talks about some other related problems, but I think the idea is clear, and how it relates to different belief attitudes. :)

That passage about teaching math is, surprisingly, equally applicable to MIT undergraduates. I was able to predict who would succeed as a math major based on whether or not they understood that the way you "read" a math  book is utterly unlike the way you read anything else. Most would-be math majors were quickly washed out by their inability to grasp this.

"Most would-be math majors were quickly washed out by their inability to grasp this."

And yet, the "educational" systems at such institutions (and I cannot pack enough scorn into those scare quotes) completely fail to recognize such deficiencies and implement instructional remedies.

Think about that for a minute.

Then ask yourself, "Do I feel as horrified about this as I *should*?"

I was, indeed, horrified and angry about it. At various points I've thought of going into math education, since I think it is disastrously bad (even, or possibly especially, at advanced levels).

I was constantly frustrated (sometimes enraged) that I absolutely could not get any of my math professors to talk about what math is and how to do it. As far as I could figure out, their intention was to make it as difficult as possible, to enhance the appearance of genius among the few people who could figure it out while the key points were deliberately obscured.

It always seemed that just as a math class was starting to get interesting, the professor would say "Now, if this were a foundations course, we'd ask the following question..." And I would be like "Yes! Yes! That's exactly what I want to know! Ask the goddamn question!" and he would say "...but of course it isn't, so instead we'll prove another boring lemma." Several times I went up after the class and said "so, that's the question I wanted answered" and the professor would look slightly scared and then I would say "so, which is the foundations class" and then they would look surprised and think for a minute and say "well, I don't think we have one, actually—maybe you could try the philosophy department."

The less said about my experience with the philosophy department the better.

So, being ein Gott verdammt Sohn von Hundin, I went and learned the foundations myself, thank you very much.

While I don't think foundations are necessarily the first thing to teach, I would have a mandatory "immigration course" for math majors, probably at the beginning of the sophomore year, that explained what math is and why we do it and how. That would get me assassinated by the priesthood, but dulce et decorum est pro mathematicam mori and all that.

Me and muflax:

me: I wonder what his reply will be...
    Seriously trying to predict here.
him: Dunno.
me: Hm, will it be a non-committal 'lol yeah' thing...
    A 'holy shit you're right' thing...
    Or even some sort of 'ha yeah but some peeps just don't have what it takes to hack it' kinda thing,..
him: I lean towards a non-committal "huh, that makes sense".
me: Yup, me too.
him: Unlikely, but maybe it gets him to read the book / put it on his todo.
    I already got other stuff on there, so there's a non-trivial chance. :)
me: Merp. ^^
him: But then, reading lists.
me: Yeaaaah...
him: Fucking reading lists, how do they work?
me: Miracles.

We shall have to recalibrate our guess-o-trons, as we failed to consider "fucking awesome" as a hypothesis. xD

You thought about going into math education just cuz it screams out to be fixed, eh?

But why think small? Muflax has come around to my idea that we should create our own university based on Theory of Instruction. (I'm calling it an "Educational University" -- a term which is not only not redundant, but actually describes something that has *never been done before*.)

We're aiming to be so obviously ludicrously superior (extremely low drop-out rates [maybe near 0%] , teaching years worth of material in months, and teaching to actual mastery as the criterion [ie, demonstrably generalized *understanding*] , all while *improving* students' mental well-being rather than stressing them crazy because *instruction will be reinforcing on a moment-to-moment level*, with equal or lower monetary costs, etc etc...), that people will spontaneously form angry mobs demanding that the professorial priesthoods repent their evil ways or be damned...

When I said muflax has come around to this idea, his exact words were:

"If it's crazy, then I'm going crazy in the same way as you. ^^ "

So, did you read the books he linked to in http://daily.muflax.com/log/10... ? (Pulled out into a list in this excerpt: https://docs.google.com/docume... )

You should look into it! Join us in our insanity! =D

How my reading list works:

There are somewhat more than a hundred paper books scattered around my living room, open to whatever page I stopped reading at. There's the same thing, digitally, in my Kindle, iPhone, and laptop. Then there are more than a dozen lists-of-books-I-absolutely-have-to-read, going back to the mid-1980s, in a collection of obscure and obsolete file formats.

I've duly added Theory of Instruction.

Good luck with your project! I utterly hope you succeed. Education is an epic disaster, and it certainly seems like it should be easier to do better.

I fear that, as with other things in which it should be easy to better (medicine, government) entrenched institutional interests are the primary obstacle, rather than technical factors. But if good, smart people don't try to do the right things, nothing will ever improve.

Alright, so muflax and I talked about it, and we both thought you should start with "Teaching Needy Kids in Our Backwards System: 42 Years of Trying" (oh god Zig, why these titles WHY? xD).

He also said to mention that you can jump into "Could John Stuart Mill Have Saved Our Schools" any time you feel like it's not clear enough what their design principles are like. (And to mention that "Mill?" is short.)

(
him: The redundant parts in "Mill?" are mostly chapter 7 and the occasional "how schools suck" parts.
    You already know that after reading "Needy Kids" or having any experience whatsoever with education.
    Also the Dewey parts are interesting, but not important and speculative anyway.
)

"Theory of Instruction" itself is... a bit of a daunting text. Well, read the preface:

http://i1112.photobucket.com/a...
http://i1112.photobucket.com/a...

(Yeah, that analogy of: "Theory of Instruction" is to education what Newton's "Principia" was to physics... Well, at least ToI isn't in Latin! xD)

I started to write more about our thoughts on how to help you break into the text, but then I realized that it makes more sense to keep this as simple as possible for now and save it for after you've finished with "Needy Kids" and "Mill?" ^^;

Put "Needy Kids" on the top of your Kindle stack post-haste! You will be very glad you did when you start to click on the material and its implications...

https://www.dropbox.com/sh/0i2...

Do it do it! xD

Now, the second part of my response, wherein I convince you of the reasons *why* you need to bump this to the top of your reading list. ^^;

You said:

"But if good, smart people don't try to do the right things, nothing will ever improve."

And:

"Good luck with your project! I utterly hope you succeed. Education is an epic disaster, and it certainly *seems* like it should be easier to do better."

Another convo excerpt with muflax:

me: "if good, smart people don't try to do the right things, nothing will ever improve."
So "Good luck with your project! I utterly hope you succeed."
XDXDXD
him: Well, dude's already trying to un-fuck-up Buddhism in the West.
And did a bunch of AI work back in the day. So yeah. Still.
"Good luck, suckers." ^^
me: It's still hilarious when juxtaposed like that. :D
him: Yes. ^^
me: But no, I mean, it's pretty much impossible at this stage for me to avoid having to make a trade-off between actually making clear how much of a Big Deal this should be, and sounding like a crank.
So
I understand.
him: Yeah. And that you should generally be cautious around "Big Deals" because of their eldritch properties.
me: Definitely true.
him: Although David's been a fan of my writing, so he must have a high crackpottery tolerance. ;)

[Although I *know* I am not in danger of going into one of Eliezer's "Happy Death Spiral" around this Big Idea; Focusing on the nitty-gritty details all the time puts the actual *structure* of its objective limitations too much in my face. xD]

So. I will now try to explain why studying the details of this science of instruction is, like, *the* most important thing you could do, in such a way that it will make sense to you now, *before* you've studied those details. xD

You said:

"I fear that, as with other things in which it should be easy to better (medicine, government) entrenched institutional interests are the primary obstacle, rather than technical factors."

Well, you're right that "entrenched institutional interests" *are* a major obstacle in general!

Buuut... three important points:

1) There is *also* a crippling lack of knowledge of those "technical factors" -- Knowledge that you *must* have to actually engineer effective (and efficient!) instruction.

Look at things like Khan Academy and Codecademy. "Entrenched institutional interests" are *not* an issue for them, and they are honestly trying to do better than the standard low standards.

And it's not that they don't have *any* success.

But once you start to really dig in to the nitty-gritty details of ToI like muflax and me, it will start to jump out at you how *awful* they are.

They are *awful*! Okay? ^^;

Summary: "Technical factors" are *really* important.

2) Yes, "entrenched institutional interests" have been, well, entrenched for a long time. They do appear very difficult to remove.

But that doesn't mean it might not be possible to remove them *very* quickly and effectively, if you could somehow find a new method that provides a position of leverage.

I assert that such a method does exist, and the leverage it will provide is *dramatic*.

But in order to really see this, you must first grasp the details of the theory.

And in order to *exploit* it, you must master the actual application of the theory to the skills of instructional engineering.

Summary: "Technical factors" are *really* important.

3) For absolutely *all* "things [that] should be easy to better (medicine, government)" --for absolutely *every* problem in the world that *could* be fixed but isn't-- there is a *single root problem* at the end of the "why chain".

What is that problem?

It is: "People are stupid".

How can you fix that?

Education.

Problems in medicine, politics, economics, science and technology, society, etc etc... The most efficient way to fix *everything* is to raise the average intelligence of the whole population. Jack up the average breadth and depth of theoretical knowledge and technical skills.

I'm not talking about "education" like instilling values and inspiration and giving-a-care and vague heuristics. Those things are also necessary, but relatively trivial.

I'm talking about taking what is now regarded as a reasonably respectable level of "genius polymath"... *and making that the new normal*.

And we actually have the theoretical knowledge, right now, of how to achieve this just through behavioral means!

Want to fix ALL the problems?

Don't try to do it yourself; Just change society so that the average person is in a *way* better position to fix things than you are now, and let them do it for you!

But in order to do that, you must first blast out those "entrenched institutional interests".

And in order to do that you must first master the engineering skills for developing optimized instruction.

And in order to do that you must first study the theoretical principles that the engineering skills come from.

Summary: "Technical factors" are *really* important.

In fact, you might even argue that they're the "primary obstacle'. ;]

Sooo... How convinced are you now? ^^;

Hmm. Many points you want addressed.

I am extremely fond of muflax because he is a charming crackpot. The "charming" is 93.2% of why I'm fond of him; crackpottery is strictly optional.

What I'm not hearing is what, practically, you want to do. Let's say that ToI is orders-of-magnitude awesome, like you believe. Then it actually doesn't much matter what you do—you'll win regardless. A plausible ToI Gnome plan is:

1) Post a YouTube video explaining how to play tiddlywinks.

2) 7 billion people hit the Like button because they are suddenly able to make tiddlywinks dance of the head of a pin. Everyone is mad for tiddlywinks. You are instantly famous and 7 billion people avidly watch all your follow-up videos.

3) By 2014, the average IQ is 237, and we save the universe from heat death.

Maybe you are not quite that confident. In that case, you need to figure out what to do next.

If you think you can do way better than Khan Academy, that's probably a good place to start, because there's already a demonstrated demand for the product.

Good luck, suckers! :-)

Damn! ^^;

Oh well, I can't say I'm *too* surprised by your reaction. I'm not sure how much of the hyperbolic caricature of my points is a joke, and how much is what you actually understood from it, though. xD

Anyway:

"If you think you can do way better than Khan Academy, that's probably a
good place to start, because there's already a demonstrated demand for
the product."

Well yeah, that's pretty much exactly what we're gonna do! :D

Or a middle stage of the plan, anyway. After it comes assembling actual physical "universities" (more like maker/hacker communities? -- for fab labs and other fun stuff).

And *before* that "replace Khan Academy" stage comes smaller scale demo courses in specific subjects. The main things we've been looking at teaching are second languages, music, and stenotype (look at this and this to see why stenotype is awesome and you want it).

At the same time, we're also going to be developing meta-ToI instruction on how use ToI to develop instruction. We know we'll need at least a few more engineers in our team to really get things done. (Even if there were no other reasons, the motivating fun of socially reinforcing each other would be worth it alone. ^^;)

So yeah, we've got a few little humps we need to bootstrap ourselves over on the way to building up a real momentum, but it's all completely *doable*, ya know?

Anyway, basically what I'm doing here in this conversation with you is calibrating on the current size of our katamari ball. xD

Just to chime in and do a bit, well, less enthusiastic propaganda. ;)

My own position on the importance of ToI is probably somewhere between "this makes learning a whole lot more pleasant because it systematically tells you how to avoid and fix the really frustrating bits" and "ok, this *could* be a Big Deal if someone actually figures out how to institutionalize it properly".

No one really knows yet *how* much more efficient or generally useful properly constructed instruction is for most cases because it simply hasn't been done. The DI folk have been able to teach children of all backgrounds very successfully, but not dramatically faster or to a higher level of competency than (lucky) high-performance children.

That still suggests to me that "pick the best-performing group of students today - that's how fast we can teach most people if we do it right" as the best possible outcome, not "everyone has IQ 200 and the universe is saved forever". Still, how important that would be in the long-run, I don't know.

I'm also totally unsure how much of an efficiency gain you'd get for more advanced skills, or adult learners. That's part of what we're trying to find out by developing ToI-compatible courses or autodidact material for all the stuff we already happen to be learning, like my language or guitar stuff. This would also include a ToI-ified instruction to ToI itself. :)

I'm still skeptical about how much I'll gain from that. So far, it seems to consistently (and with relatively little effort) eliminate the "gah, how does that *work*?!" parts, but overall, I don't see a massive efficiency boost yet. So I can't yet learn in 3 weeks what would previously take me 3 years, but maybe that's just because we're not yet skilled enough at ToI. Only one way to find out, though.

Also, that tiddlywinks thing *is* actually kinda the unnofficial plan. "Write a meta-ToI course, develop some demo courses to show how awesome this is, everyone learns ToI and instruction sucks no more."

(And now I see Owen has ninja'd me to the last point, but whatever.)

Well, again, seriously, I think this is interesting and I hope you win big. I don't know why you want to persuade me to be more enthusiastic than that; I'm some random guy. Put something enticing on reddit if you want to gauge interest.

The approach of developing small demonstrations seems like the right one. If those do work well, you will gain momentum automatically. The scenario is inherently exciting, and lots of people will jump on board.

There's a ton of music and second-language self-instruction materials out there. I know from painful experience that 99.9% of it sucks totally, but—to contradict my own previous suggestion—it might be hard to stand out in the middle of all the rubbish. Perhaps it would be better to find something for which self-instructional material is rare? (Maybe that's part of why Khan Academy won big—there was much less teach-yourself-math stuff available.)

Have you, by any chance, read "Guitar Zero" or David Byrne's "How Music Works"? Both are on my Kindle unread, so I'm curious if you have an opinion about whether I ought to move them up the queue.

I have a post coming up about how I realized, after writing the first draft of my novel, that maybe I ought to read something about how to write a novel. Man, if you think how-to-play-guitar books are crap, you should check out how-to-write-a-novel books. Epic.

The only good advice I got was "leave sex scenes and comedy to professionals." The next instalment of my novel violates that thoroughly, however, which proves that I can't learn from good advice.

Your previous suggestion was right. Standing out from the middle of all the rubbish is *exactly* what we want to do. Michel Thomas provided a proof-of-principle for that strategy...

Yeah, I don't really expect you to be convinced at this point. Mainly, it's just that seeing exactly *how* things don't get across properly gives me ideas... Making a public broadcast on reddit or something is scarey;  I'll stick to feeling things out with individuals for now.

(I don't expect you be convinced that it's a Big Deal by anything I say here, that is. You might be convinced to start reading the books, though, in which case you might end up being sucked in. xD)

Yeah, I've heard of that book. Haven't read it. Muflax said, "From what I can tell, Guitar Zero is mostly just pro-immersion propaganda."

And then there was this from this wiki page:

"... a beginner learns guitar by learning notes and chords,[36] and irregularities make learning the guitar difficult[37]—even more difficult than learning the formation of plural nouns in German, according to Gary Marcus.[38]"

If Gary Marcus had read ToI, he would realize that screams out: "This concept is an instance of the 'transformation' logical structure. Do a subtype analysis. Apply John Stuart Mill's "Method of Residues". Duh."

But I'm sure that, for him, "How to teach German plurals" is just this inherently mysterious question to which the answer is some vague "practice? immersion?" kinda thing.

Actually, "develop instruction for German nouns (cases, plurals, adjective agreement etc)" is already on my project list. I should bump up its priority a bit...

(I think Michel Thomas's German course was his least complete one (although I haven't done the Italian, I'd guess it to pretty much follow the Spanish). In German, he covered syntax and verbal inflections pretty well, but he almost completely left out any *mention* of the noun complexities. I suspect he had probably taught those in his private tutoring sessions in the past, but skipped them when he made the recordings --each course was recorded over a weekend, I believe-- simply because he ran out of time.)

I've read "Guitar Zero" and found it  pretty much useless except for the "adults aren't magically incapable of learning stuff" and "you gotta actually practice" propaganda, which is important and way too under-represented, but which you're probably already aware of. :)

(I also try to include those points whenever I can, especially because of Khatzu whose style I stole^W was influenced by, so I certainly approve, but at least I found the book redundant.)

"How Music Works" is also on my reading list, so maybe I'll comment on it soon.

(I've avoided "how to fiction" material so far because it never covers the genres I care about, although I found "how to make a compelling RPG campaign" stuff like this interesting.)

 Here's one concrete example from "War Against the Schools' Academic Child Abuse", page 86 (scanned page attached).

"
... studies were conducted using the videodisc programs. My favorite compared the performance of "learning disability" high school students and "remedial students' who had failed previous science courses to the performance of advanced-placement students who were in their second year of chemistry. This study was done as part of the initial field-testing of our videodisc program on Chemistry and Energy. The learning-disability kids and remedial kids went through the videodisc sequence, after which they and the advanced-placement kids were tested on bonding, equilibrium, energy of activation, catalysts, atomic structure, and basic properties of organic compounds. Although the advanced placement students were light-years ahead of the video group in achievement (close to the 90th percentile in math and science), the remedial students outperformed them on every topic, and even the learning-disabled students outperformed them on bonding and equilibrium. The mean post-test scores for the video group was 75, compared to 71 for the advanced-placement students.
"

[This is the study, I think (fucking paywalls). Tell ya when we get it.]

Okay, now note:

- This study was done as "as part of the *initial field-testing*" of the videodisc program. It wasn't even a product at a super mature stage of development.

- They did this with "videodiscs". It's quite possible that they were forced to cut corners a few times due to the difficulty of getting all the desired effects with this then-new technology.

- And yet we still see the learning disabled kids doing not just "just as well", but actually a bit *better* better than the elite Advanced Placement kids on *everything* the instruction covered.

- Also note that this was an extremely *recent* (to the time of testing) and very *short-term* intervention!

What would happen if these "learning disabled" kids had been taught through such methods since the very beginning of their education? How much further might they go? What about AP kids?

Basically what I am saying is, I think the enthusiasm is warranted.

Thanks for the "Guitar Zero" advice.

That chemistry videodisk story is in "extraordinary claims require extraordinary proof" territory.

Probably a lot of the people who take AP chemistry are pre-premed: children of doctors who are more interested in aceing the test than they are in chemistry. That's good, for you. They have the money and motivation to buy better instructional material.

If you can make AP chemistry videos that do what the book claims, you can make more money than you can imagine, take over the world, and save the universe.

Go for it!

(Why hasn't someone already done that? "Engelmann's too nice" doesn't count as an answer.)

(I'm currently trying to get the original paper, and I share your skepticism about this particular demonstration. The "we taught them up to AP level" claim I believe, but "we did even better" I'd like to see more evidence for first.)

That's what I'm saying! Why don't you go after the elites?! Or at least people who actually have money and/or influence?

Michel Thomas did that, of course. We're somewhat skeptical how much of his own method he really understood, but regardless, he kept it all a secret, tutored Hollywood folks and made a decent living with it. And yet crap like Rosetta Stone still exists.

Zig and his crew focus primarily on poor kids for political reasons. They tried to get their stuff adopted by the school system and failed miserable. They considered focusing on high-performance students, but decided those already got enough of an advantage and you really need to lift up the disadvantaged.

That's nice, I guess, but the success so far has been fairly limited (although all the schools they *do* run are excellent at what they do). So yes, the reason almost all their stuff is K-5 is because of their specific ideology and a massive fail of strategic thinking.

Especially because it doesn't actually take much to run a DI school! The courses are entirely scripted and only designed once. Educating the teachers how to use the courses is also not particularly difficult. The main obstacle is to find a place that actually let's you run the course. (The amount of sabotage on every level they've faced is impressive.) So running private DI schools should be fairly trivial and comparatively cheap.

So yeah, they're really running a Nice Charity For Cute Little Puppies, and that's why they don't have any significant funding.

And here's the paper.

tl;dr: The "we did better than AP students" claim is not justified by the results and the tiny sample size. The result is cool in terms of improving remedial students, but not more. More research is needed(tm).

(It is also not true that they matched or outperformed them at every topic, but I take it that this was an honest mistake.)

So, without having actually read any of this stuff, here's a hypothesis. Let's say, as a first order model, teaching = lectures + interaction. Let's say that disadvantaged kids get lousy teachers (because, racist evil, or whatever). Let's say most lousy teachers are lousy at lectures and lousy at interaction. They are totally clueless and can't explain anything, or not motivated to.

So what the studies actually show is that if you force lousy teachers to deliver a canned, high quality lecture verbatim,  students do a whole lot better. That raises the lecture component well above average. For the stuff these guys looked at, the interaction component was maybe much less than 50% of teaching, so maybe the overall quality of teaching goes above average in canned-lecture condition.

This would be an important result if true, and I find it perfectly plausible. So then what is implied? Well, at minimum that you should force lousy teachers to deliver canned lectures. Presumably they resist that for economic reasons, which is why it hasn't happened.

Plausibly, you can replace the lecturing component of teachers with a recorded presentation. (It might be important to hear it coming out of a live monkey's mouth, because evopsych. But probably not.) Obviously teachers will strongly resist that (and cite "interaction is critical" as a justification). However, from what I'm reading in Google News, there's a widespread sense that this is about to happen regardless, because live monkeys are expensive and Khan Academy.

Where does that leave ToI and you?

Well, offhand, maybe ToI is a recipe for designing the high-quality lectures. Maybe excellent teachers only roughly conform to ToI, intuitively, and they would be more excellent if they knew it explicitly. So then the product is that meta-knowledge.

Maybe ToI is a magic bullet for the interactive component, which is harder to automate. Maybe teachers should all be forced to use it. Maybe you can explain how to do that. Maybe you should automate the interactive component, and then we can get rid of teachers altogether.

My parents were high school teachers, btw. They talked about teaching and how it works around the dinner table.

DI's stance is that teachers are bad instructional designers, and it's just stupid to require them to do this job. Highly rigid courses are a compromise to make it possible for bad teachers to deliver good lessons. Competent teachers need less scripting.

Thing is, they did a lot of testing with that, and found that even highly competent teachers would screw up instruction from time to time if they were forced to improvise. Having a rigid script helps them too. (And competent teachers are very rare.)

Your two-component model is basically correct. Any good instruction must follow the principles in ToI, and any good interaction must obey the standard insights of behaviorism. You need to have both working to get good education, and teachers are very bad at either.

In DI's ideal world, teachers wouldn't do instructional design and always just follow perfect courses. Their actual job would be the equivalent of a dog trainer - to present the instructions, test learners and hand out reinforcers. That's plenty hard already, but within the reach of most actual teachers we have now.

Most of that can be automated, and probably should be. Given current technology, the only cases where you *need* actual physical teachers is when your learners don't have the meta-skills to follow your automated courses (e.g. they can't read yet), or when you try to teach them something they don't want to learn and have no use for. Which you probably shouldn't.

(I suspect the primary reasons students currently hate a subject is
because the instructions are horrible and frustrating, not because they
dislike learning about the topic or because it's inherently hard.)

ToI tells you how to design the courses and solve the "lecture" part. Behaviorists (and animal trainers) already solved most of the "interaction" part ages ago, and DI just integrated their insights (and cooperated with lots of behaviorists).

(This was also borne out by Project Follow-Through, the big educational study that compared a lot of teaching methodologies. Only the two behaviorist approaches (DI and one other) improved children's behavior, and only DI actually taught them something.)

I suspect the primary reasons students currently hate a subject is
because the instructions are horrible and frustrating, not because they
dislike learning about the topic or because it's inherently hard.

Agreed.