Edited based on the outline kindly provided by Gram_Stone, whom I thank.
There is a skill of reading and thinking which I haven't learned so far: of looking for implications as one goes through the book, simply putting it back on shelf until one's mind has run out of the inferences, perhaps writing them down. I think it would be easier to do with books that [have pictures]
- invite an attitude (like cooking shows or Darwin's travel accounts or Feynman's biography: it doesn't have to be "personal"),
- are/have been regularly needed (ideally belong to you so you can make notes on the margins),
- are either outdated (so you "take it with a grain of salt" and have the option of looking for a current opinion) or very new,
- are not highly specialized,
- are well-structured, preferably into one- to a-few-pages-long chapters,
- allow reading those chapters out of order*,
- (make you) recognize that you do not need this knowledge for its own sake,
- can be shared, or at least shown to other people, and talked about, etc. (Although I keep imagining picture albums when I read the list, so maybe I missed something.)
These features are what attracts me to an amateur-level Russian plant identification text of the 1948.** It was clearly written, and didn't contain many species of plants that the author considered to be easily grouped with others for practical purposes. It annoyed me when I expected the book to hold certain information that it didn't (a starting point - I have to notice something to want to think). This is merely speculation, but I suspect that the author omitted many of the species that they did because the book was intended to convey agricultural knowledge of great economic importance to the Soviet population of the time (although some included details were clearly of less import, botanists know that random bits trivia might help recognizing the plant in the field, which established a feeling of kinship - the realisation that the author's goal was to teach how to use the book, and how to get by without it on hand). I found the book far more entertaining to read when I realized that I would have to evaluate it in this context, even though one might think that this would actually make it more difficult to read. I was surprised that something as simple as glancing at a note on beetroot production rates could make me do more cognitive work than any cheap trick that I'd ever seen a pedagogical author try to perform purposefully.
There may be other ways that books could be written to spontaneously cause independent thought in their audiences. Perhaps we can do this on purpose. Or perhaps the practice of making inferences beyond what is obviously stated in books can be trained.
* which might be less useful for people learning about math.
** Ф. Нейштадт. Определитель растений. - Учпедгиз, 1948. - 476 с. An identification key gives you an algorithm, a branching path which must end with a Latin name, which makes using it leisurely a kind of game. If you cannot find what you see, then either you've made a mistake or it isn't there.
The experience of reading this article is extremely disfluent. In particular, it wasn't immediately clear to me that the quoted sentences after the book excerpts were inferences about the previously quoted sentences. Instead, I read the quoted sentences independently of any context, which instilled a great sense of confusion.
However, I find this valuable. I'm going to paraphrase your anecdote so that maybe this article will get some of the votes that it deserves. Please tell me if this is accurate:
Thank youso much! It's my fault that I miscommunicate things; I will correct it when not on the phone. (And look up some of the words you use). 05.05. Edited the text; thank you!