Actually I'd argue that a theory of composition of Chinese is somewhat overrated. It really only makes sense for the common list of Chinese radicals. It may be

Actually I'd argue that a theory of composition of Chinese is somewhat overrated. It really only makes sense for the common list of Chinese radicals. It may be useful as a device for remembering characters, but the explanations each person comes up with are bound to be idiosyncratic. For example, you seem to be implying a connection between the meanings of 王 (king) and 上 (up) based on their graphical similarity. This is purely an artifact of regular script, one of several different scripts for Chinese. This similarity goes away entirely in older scripts and cursive scripts. In general graphical analysis of modern Chinese script is usually extremely misleading and to do accurately often requires an extraordinarily convoluted process. I would not recommend it as a way of formally analyzing Chinese. Through the repeated process of phonetic borrowing (taking a character similar in sound to an unrelated word and using it to represent that word as well); changes in Chinese script over the millennia; variant characters that have been consolidated, split up, and re-consolidated; variant meanings that have been consolidated into a single character, split up into separate characters, and re-consolidated; erroneous interpretations of character structure between different iterations of Chinese scripts; and normal semantic drift of words over time, it is often maddening to divine the meaning of Chinese characters from their graphical forms that is any more specific than the usual context offered by a Chinese radical. This also means that ancient Chinese doesn't have too many graphical puns. There are isolated examples e.g. where these puns are only used to get around prohibitions on certain characters (e.g. characters used in an emperor's name could not be written for the duration of that emperor's rule, so graphical puns would occasionally be employed). This persists in modern Chinese where to get around some censorship, graphical puns are employed. Likewise, I am very skeptical of viewing the poem as an intralingual rosetta stone for learners of the writing. Just writing it in a different script changes a lot of the graphical structure! There are no commentators I know of, Chinese or otherwise, that even attempt to analyze the graphical structure of the characters in the poem. Graphical analyses really only have value for the earliest stages of Chinese writing (oracle bone script and to some degree bronze script). And even then things are painful, especially because the idea that a word is identifiable with its character starts to fall apart. You start having to think harder about whether a single character is meant to stand-in for several words or whether multiple characters all can be used for a single word.