But perhaps Cangjie really made some contributions in the existing Chinese writing system: instead of the inventor, he might be a collector and collator of scattered Chinese characters in ancient China.Thanks to many a contributor like Cangjie and the common people using and spreading characters, a complete well-developed writing system had finally come to birth. The indisputably evidence is Chinese character inscriptions found on turtle shells dating back to the Shang dynasty (1766-1123 BC), formally called Oracle bone script. Of the 4,600 known Oracle bone logographs, about 1,000 can be identified with later Chinese characters, and the other unidentifiable ones are mostly the names of people, places or clans.

In view of formation, written Chinese is a script of ideograms. Xu Shen, in the Eastern Han Dynasty (121 AD), was a distinguished scholar who had attained unparalleled fame for his etymological dictionary entitled Shuo Wen Jie Zi, whose literal meaning is "explaining written language and parsing words". In Shuo wen, Chinese characters are classified into six categories, namely pictogram, ideograph, logical aggregates, pictophonetic compounds, borrowing and associate transformation. However, the last twos are often omitted, for the characters of these categories have been created before but somehow borrowed to represent another meaning, or detached into separate words. Generally, Chinese characters fall into four categories in view of their origin.

Pictograms (Xiang4 xing2 zi4) Pictograms are the earliest characters to create, and they usually reflect the shape of physical objects. Examples include the sun, the moon, a woman, fire. From this picture-drawing method, the other character forming principles were subsequently developed. Over a long history, pictograms have evolved from irregular drawing into a definite form, most simplified by losing certain strokes to make ease of writing. Therefore, to see the actual picture of what it represents, you must have a lot of imagination as well as knowledge of the origin of the character and its evolution. However, only a very small portion of Chinese characters falls into this category, not more than 5 percent.

Ideograph (Zhi3 shi4 zi4)
Also called a simple indicative, Ideograph usually describes an abstract concept. It's a combination of indicators, or adds an indicator to a pictograph. For example, a short horizontal bar on top of a circular arc represents an idea of up or on top of. Another example: placing an indicative horizontal bar at the lower part of a pictogram for wood, makes an ideograph for "root". Like pictograms, the number of this category is also small, less than 2 percent.

Logical aggregates (Hui4 yi4 zi1)
It is a combination of pictograms to represent a meaning, rather like telling a little story. A pictograph for person on the left with a pictogram for wood on the right makes a aggregate for "rest". This story-telling formation is relatively easier to learn, yet most of aggregates have been reformed into phonetic compounds, or just replaced by them.

Pictophonetic compounds (Xing2 sheng1 zi4)
Also called semantic-phonetic compounds, just as the name implies, it combines a semantic element with a phonetic element, taking the meaning from one and the phonetics from the other. For instance, the character for ocean with a pronunciation of yang2 is a combination of a semantic classifier which means “water” with the phonetic component yang2, referring to goat or sheep on its own. This last group of characters is the largest in modern Chinese, making up around 90% of all Chinese characters. The superiority of phonetic-compounds over the first three categories lies in its unique phonetic components, for many an object and concept are hard to express through photographs or ideograms, and its association with the character pronunciation helps Chinese vocabulary extends much faster than logical aggregates. Therefore, most newly created characters take this more scientific formation approach.

However, over the centuries evolution, the Chinese language has undertaken such a great change, that most pictophonetic compounds don't pronounce as its phonetic elements any longer, and the semantic components appear even