I've never seen any really good expositions of elementary mathematics (middle school or earlier). A good college-level textbook, written for people with an interest in mathematics, reads like a novel or an impassioned essay. The most famous example I can think off the top of my head of is Spivak's Calculus - the author has a vision for the subject, and wants to communicate their passion to the student.

As an example, the rule for multiplying fractions together takes some quite clever and pretty arguments to prove, but in math class and in classroom textbooks, this is just presented as:

Fraction multiplication rule: Multiply the numerators and the denominators together separately.

Looking at the difference in presentation, you'd almost think elementary arithmetic and geometry aren't part of mathematics. Where are the proofs? Is there even a single book in existence that proves the above fact at all?

Are there any books about elementary math written like actual math books? Note that I am not looking for things like "The Number Devil", which (to be fair I haven't read it) you probably wouldn't mistake for an actual textbook. I'm looking for systematic expositions of basic math, covering (and proving!) all the theorems, rules of computation, and definitions covered in elementary to high school (not necessarily all in one book - that's a lot of material).

I don't necessarily want it to be readable by 8 year olds, that might be asking a lot, but it would be nice to have something you could recommend to a bright 12 year old to give them a more solid foundation and inspire more of an interest in math. Something well written, that wants to make the subject interesting and beautiful, is preferred to a dry going-through-the-motions proof compendium, but of course that's basically like saying "make sure the book is actually good".

-
@paulgarrett From a strictly logical perspective, the rationals are very often defined as equivalence classes of $\mathbb{Z} \times \mathbb{Z}/\{0\}$, with $(a_1,b_1) \cong (a_2,b_2)$ if $a_1b_2=a_2b_1$. Then multiplication of fractions is defined by $(a,b)(c,d) \cong (ac,bd)$. What needs to be proven is that this definition is even well defined. –  Steven Gubkin Apr 22 at 13:12
@StevenGubkin, I am well aware than "X is often defined as Y", but that is a separate question from the actual mathematical function of either. Also, the possibility of writing things in first-order logic or other formalities does not lend anything greater gravitas or correctness. The function of the rationals is as field of fractions of the commutative ring $\mathbb Z$, and that requires/entails certain properties (apart from any notation). The construction as equivalence classes of somehow-denoted pairs of integers, and notations about that, are secondary to the function. –  paul garrett Apr 22 at 13:28
@StevenGubkin This is an interesting debate, but it's probably better suited for the chat! –  Jack M Apr 22 at 13:34
@StevenGubkin The StackExchange button at top left - it's in the drop down. You can ping people using @ once you're in there. I'll edit a proof into the post if you do think it's important, though. –  Jack M Apr 22 at 13:39
@paulgarrett can you join me in the chatroom for this site? chat.stackexchange.com/rooms/13591/mathematics-educators –  Steven Gubkin Apr 22 at 13:48

I think you'll find some of what you want on Berkley mathematician H.H. Wu's homepage.

More precisely, see: Pre-Algebra (pdf) and Introduction to School Algebra (pdf).

Note: I mentioned the same homepage (and the two pdf textbooks) in an earlier MESE post here; I would have just re-posted this as a comment, but I believe it is the actual answer to your question (!). I also ought not mark your question as a duplicate of the one I answered earlier, since they are quite different.

-

Below are three books for what I think you want. At one time these books were all very well known, at least in the U.S., probably because virtually every public library (even small town libraries) used to have copies of these books.