Nate Solon has just provoked with a blog on Lichess:
Why Chess Books Don't Work

He has some point -- I mean, if books made you a better chess player, I should be world champion by now...

But I commented, crossly:

To the extent that Chessable has a theory of learning, it's a very impoverished one.

I believe Pillsbury was once challenged to remember a list of obscure words, starting Antiphlogistine, Madjesoomalops etc
Spaced repetition/recall seems perfect for this task where any meaning in the content is irrelevant.

A book will have, as you suggest, an implicit theory of learning, but I suspect will be richer than you give it credit for, perhaps having several on one page. I don't know if you know the book Winning Chess by Chernev and Reinfeld -- I suspect both authors would struggle to articulate a 'theory of learning' of the sort you seem to think is required, but what a great pair of teachers.

I know many books from Everyman which use the Socratic method, of presenting a question before giving an explanation.
Others engage with wit, colourful language, striking examples, slogans, scaffolded exercises etc. Implicit maybe but Ford save us all from learning opening lines by rote.

Good-quality verbal explanations in books can unlock a whole opening system for you, as can a good example.

The right book at the right time can produce a step-wise improvement -- I can remember two or three from my early playing days (Winning Chess Chernev/Reinfeld, Middle Game I & II Euwe/Kramer, Secrets of Practical Chess Nunn).

Perhaps the reason most books don't 'work' is because most books are not good books -- or not good for you at the time.

I can remember acquiring, on Silman's recommendation, the book on Combinations by Znosko-Borovsky. I learned nothing from it, as Chernev/Reinfeld had done their job well. I also learned next to nothing from Fine's Basic Chess Endings -- it was far too advanced for me when I picked it up.