A small number of characters began as pictures. Most everyday characters are compounds. If you only remember the classroom story that Chinese writing is drawings of suns and moons, you will misunderstand both history and software. Pictures run out as soon as you need grammar words, names, and abstract verbs. The writing system grew other methods.
The usual modern pattern is a semantic part plus a phonetic part. 河 uses the water component plus a sound clue. 湖 does the same job with a different phonetic. Once you see that pattern, a page of Chinese looks less like thousands of unique sketches and more like a large set of reusable parts.
The traditional toolkit, used carefully
Teachers still mention the six-fold classification (六书). It is a historical toolkit, not a laboratory procedure you can run on every Unicode character. The useful everyday buckets are:
Pictographs. Graphs that began as pictures of objects. Pictographs covers the limits of this class. 日, 月, 木 and 马 / 馬 are the usual textbook set. Even these have been squared by regular script. They are not photographs.
Simple ideographs and combined ideographs. Marks that point at an idea: 上 and 下, or 休 as a person beside a tree. These explanations are teaching stories. Some are historically tidy. Some are later mnemonic.
Phono-semantic compounds. The majority of the inventory. One part hints at meaning, one part hinted at sound at the time the character was formed. Sounds have shifted for two thousand years, so the phonetic is often only a cousin of the modern reading. That is why learners say "the sound hint is wrong." It is old, not random.
Loans and other leftovers. Some graphs were borrowed for a similar-sounding word, then the original meaning got a new graph. Specialist books argue over individual cases. You do not need those arguments to convert a file.
Radicals are how dictionaries name one of the parts, usually the semantic one, for lookup. They are not a rival theory.
Why most characters are compounds
Spoken language has more words than a person can draw. Chinese writing kept a morpheme-sized graph for many words instead of switching to an alphabet. To stay learnable enough for clerks, the graphs had to share parts. Phonetic compounding is the productivity trick. Once 青 is available as a sound clue, a family of characters can be built around it.
This is also why stroke count and "looking like the thing" are poor guides to difficulty. A compound can be visually busy and still easier to remember than an arbitrary picture, because you already know the pieces.
Oracle bone script already shows compounding, not only pictures. The modern page is the same idea after clerical script, regular script, printing, and in some regions simplification.
Why this matters for conversion
Simplified forms sometimes keep the phonetic and drop strokes in the semantic part, or the reverse. Sometimes they merge two compounds that used to be distinct. That is one reason a one-to-one mental map fails. You cannot reliably invent the Traditional form by "adding the missing strokes" from first principles. Dictionaries do the mapping instead.
It also explains a proofreading instinct. If a converted character has lost a water component in a word that is clearly about rivers, you look twice. If a name picked 後 instead of 后, you look twice for a different reason: a merger, not a missing pictograph. See one-to-many mappings.
OpenCC does not parse semantic and phonetic parts as a linguist would. It matches whole characters and phrases. Structure is for your understanding, not for the engine.
How to study structure without drowning
- Learn a few honest pictographs so the textbook story has examples.
- Learn common semantic components (the usual radicals).
- Notice phonetic series when they still help, and ignore them when they do not.
- Do not delay reading real sentences until you can classify every graph.
For living standards, return to Simplified vs Traditional Chinese. For the converter, structure is background knowledge. The job is still: pick a target, convert, proofread.
Frequently asked questions
Is Chinese a pictographic language? No. It is a character language with a small pictographic minority and a large compound majority.
If I learn all the radicals, can I read any character? You can often guess a field of meaning and sometimes a sound. You still have to learn the word.
Why do phonetic parts not match Pinyin? Because the character was formed when the pronunciation was different, or because the phonetic was approximate even then.
Do Simplified characters use a different construction system? No. They use the same compounding idea with a different official inventory.
Does character structure affect Pinyin tools? Only indirectly. Pinyin tools read modern pronunciations, not the ancient phonetic hint.