Why books don't work (2019)

GP didn't specify, but I assume they meant the global West in general and not just the West-Coast.

I too am underwhelmed by the article, but I don't think it's fair to call the books he describes "mediocre pop science books". By way of reminder, they're Dawkins's The selfish gene, Kahneman's Thinking, fast and slow, and Diamond's Guns, germs and steel.

These are definitely pop science books. Like any pop science books, they can reasonably be accused of simplifying their material, favouring a particular interpretation of science that can be looked at in different ways, etc. The underlying science is (like all science) fallible too. But all this is just to say that they're pop science books. That doesn't make them mediocre pop science books, and it seems to me hard to deny that they are all unusually good pop science books.

It's not relevant for this purpose whether all the science turns out to be correct, nor whether Dawkins & Kahneman & Diamond let their own opinions obtrude too much; Matuschak's complaint isn't "I learned all this stuff and some of it turned out to be wrong", it's "I read these books and it turned out I'd learned less than I thought".

It is relevant whether the books present a nontrivial amount of content, whether they're explained clearly and engagingly, whether the ideas in them are (or were when the books were written) at least somewhat unfamiliar to nonspecialist readers. On all those counts, all three of these books do well.

So I don't think he'd have got better results by reading other less-mediocre popular science books. I also don't think the results would have been any better if he'd done the same thing with books on (say) philosophy, or history, or language. The problem is with some combination of (1) what he did and (2) what he hoped to gain from it, not with the particular books he chose.

I do agree that there is a big problem with (1) and (2). If you actually want to learn something then you need to do more than read one book on it; and reading a good book (even if it's what you call a mediocre popular science book) once will do you some good but what it does for you won't be to enable you to expound in detail on its contents at parties.

Oh, books do work. Take, for example, early reading. No single book will teach a child how to read. One may reinforce letters, or a more advanced one may reinforce words, or later how they are used in sentences, then vocabulary, and so on until a point when the child has developed the skill to read unassisted. One could argue that the entire fiction genre is merely a roundabout way of reinforcing vocabulary (if one is inclined to do a cynical utilitarian take on the knowledge value they provide).

Extending from that, it seems a bit presumptuous to assume that a single book might then be able to convey all there is to any topic due to its size (compared to, say, listicles), given that reading itself is one of the most basic skills and one that - as we've seen - cannot generally be taught with a single book.

What books can do is expose you to some idea, in some form, so that you can soak up something from it and adopt it into your own mental framework. Pop culture / self-help books will often try to "sell" a central idea by showering you with stories and anecdotes (e.g. the 1-2-3 system in parenting books), but if you have the presence of mind to already understand that the examples are merely an attempt to add weight to an argument, then you can quickly summarize the meat (the "1-2-3", in this case) and think for yourself whether/when to apply that technique, vs other techniques that you might learn elsewhere.

There's a slippery slope in terms of how many core ideas one may pack in a book, vs how many supporting arguments one may add around the core idea. Add too many core ideas and you get something like a dry university textbook. Add too little, and you get books like Malcolm Gladwell. Mind you, having a lot of "frills" doesn't necessarily make a book bad. For example, many consider Godel, Escher, Bach to be a great read, even though [spoiler] it can be promptly summarized as an exploration on occurrences of recursiveness (through very colorful allegory). The beauty of GEB is that the supporting stories themselves give you a glimpse of new domains to explore.

At the end of the day, just like a library is a collection of disparate nuggets of human knowledge and no one's expected to uniformly learn all of it, books are but pieces of an ecosystem, with various degrees of "completeness" (as far as their core subjects go) and various degrees of flair; what one chooses to read or skim, and what they take away individually from each book is what the relationship between humans and books are all about.

A compelling quote, but alas: https://quoteinvestigator.com/2016/06/20/books/

I often find myself really bad at remembering what I read in a book or blog post or heard in a lecture on youtube or Coursera beyond the larger point, so I'd agree with the author at least that far.

However, when a question is raised relating to something I read, eor a point is stated that conflicts with something that I read, I usually find I can remember where I read about it. Even if I didn't internalize the idea I can look it up pretty easily, and 90% of the time being able to look up something is good enough. So I think whether books work depends on your goal: is it to perfectly internalize something, or to be able to have something you know you can refer to later?