Hypothesis: Hanzi hesitated between transforming into a purely phonetic writing system between 1500 and 3000 years ago but retained its approach, which is based on pictographs which then are formed into phonemes.
My hypothesis is that we should try to read characters phonetically from right to left and from top to bottom; i also think we should look for finals rather than initials first, but that we should look for both initials and finals when searching for phonetic clues.
I do not think we should look at any phonetic choice as arbitrary. The logic of the choice may however be dependant on historic/dynastic conditions which are entirely irrelevant today! That, plus language drift, explains why some believe that phonetic character choicses are sometimes arbitrary.
I don’t think we can develop rules like we have in western languages as to proper orthography. “read this way, not that way”. When looking at a character we should first
1) Isolate phonetic element(s)
2) Isolate semantic element(s)
3) Use pictographic elements as a “fallback” “catchall”.
4) Try to figure out if/how semantic and phonetic elements are mutually reinforcing “not only does this MEAN x it ALSO sounds like y!”
Pictographics are a supplement for advanced characters but the base for basic characters.
My favorite picto-phonetic character is flag. It’s the man-in-stocks on the left and a tool on the right with a dashed-line on top. It’s the soldier (not freeman) on the left and his tool on the right is that god damned flag which is the top line (I believe the English word for top strips of a flag mounting is: penant). The top line (penant) is clearly pictographic. The other two elements are phono semantic but also pictographic.
Fang+qi+zhu-yi=Flag. Sounds like Fahne, which is German for flag. Looks like a flag-man.