Arbitrary sign

Qiú defines sign as follows: “Graphic symbols ... without any relationship either on the semantic or phonetic level are signs.” (Qiú, 2000, p. 15) For clarity I use arbitrary sign.


See signific.

Empty component

A “component [that] provides an empty form, which is not related to its full form, meaning, or sound” (Outlier). For example, 美 “beautiful” originally depicted a person wearing a headdress. The modern form 美 consists of the of top of sheep 羊 and 大. While these components have meaning on their own, in 美 their meanings are irrelevant. See also standardization.


“A loangraph is a homophonous or nearly homophonous graph borrowed to write another word.” (Qiú, 2000, p. 261) A somewhat analogous example in English would be to use the sign “2” to write “to”(“will go 2 U”).

Mnemonics and reinterpretations


From Wikipedia:

A mnemonic (/nəˈmɒnɪk/, the first “m” is silent) device, or memory device, is any learning technique that aids information retention or retrieval (remembering) in the human memory. Mnemonics make use of elaborative encoding, retrieval cues, and imagery as specific tools to encode any given information in a way that allows for efficient storage and retrieval. Mnemonics aid original information in becoming associated with something more accessible or meaningful—which, in turn, provides better retention of the information.

In the context of Chinese and Japanese characters mnemonics usually work by analysing the elements of a given graph in such a way that a short story can constructed. When the graph is a semantic compound, such an analysis may correspond to the original composition of the graph (for example “three trees 森 equal a forest”).

However, most graphs are phono-semantic; a number of frequent graphs are originally pictographic—but may be have become arbitrary signs through standardization; only very few are semantic compounds (and a lot of them have become unintelligible in their modern form as well).

In practice mnemonics are based on a reinterpretation of the (elements) of a graph and building a (highly subjective) memorable story.


Reinterpreting graphs is almost as old as the graphs themselves are. A lot of the analyses of Xǔ Shèn (ca. 58–ca. 148 CE) amount to reinterpretations, because the origin of a lot of graphs was already lost at that time. People have been reinterpreting and telling stories about graphs ever since. Occasionally such stories may be so compelling that they get firmly attached to a given graph. Mostly however, those stories are as subjective and arbitrary as any mnemonic made up by a student.

Phonetic element

Qiú introduces phonetics that are part of compound characters as follows:

The phonetics of phonograms are also phonetic symbols. There are two kinds of phonetics. One kind is borrowed purely for the purpose of expressing sound, e.g., the phonetic “化” huà of the character “花” huā. Another kind also has a semantic relationship with the word represented by the phonogram. For example, a type of ear ornament made from jade is called {珥} ěr (homophonous with the character “耳” ěr “ear”); the character 珥 ěr consists of “玉,” “jade” plus “耳” ěr (玉, when used as a component on the left side of a character is written “⺩”). The component “耳” ěr is a phonetic bearing a semantic relationship to 珥 ěr. Phonetics of this type can be viewed as phonetic symbols which are concurrently semantic symbols. (Qiú, 2000, p. 17)

Phonetic elements in Chinese are almost always loaned phonetics (ibid). Perhaps an example in English could be the following. One could write tonight as 2night. The graph “2” (normally used to write the word “two”) is in that case used as a loaned phonetic. That “2” also has a meaning is ignored in spelling 2night (just like the meaning of “化” is ignored in “花”above). On the other hand, the alternative way to write threesome, using the graph “3” in 3some, would be an example of the use of “3” as a loaned phonetic that gives both a pronunciation (its phonetic aspect) and meaning (its semantic aspect).


Ochiai explains protoform (初文) as: “When the structure of a graph later on has changed, [we] call the earlier form of the character the protoform.” (Ochiai, 2016, 用語解説 p. 95)

Semantic compound

A semantic compound combines meaningful elements that are not being used for their phonetic value.

This term I use as the equivalent for Japanese kaii moji 会意文字, Chinese huìyìzì 会意字. Qiú uses syssemantograph, which I find a horrible word. Elsewhere it’s often compound ideograph, but I agree with those scholars that think that “ideograph” and “ideogram” can be confusing.


The part of the graph that is specifically added to classify a graph semantically, to determine approximately what kind of meaning the word might have that the graph points to. Same term as determinative.

For example “亻” (derived from 人 as “person”) is often added to words that have to do with human affairs. The graph 他 pertains to other people, or the personal pronoun he/she/it; the graph 仕 can points an official; the graph 侍 points to a servant; etc.

In the past the term “radical,” might have been used in this sense. That usage was unfortunate, because strictly the term “radical” has to do with the “root” or “origin” of a word, which is not what a signific is about. In fact, the other part of the graph is much more informative in that regard, because it points directly to the word in question, while the signific merely puts the word in a category. For that reason it’s preferable to use the term signific or the term determinative, and avoid using “radical.”


The process in with unique or difficult to draw shapes are redrawn with easier or familiar shapes, often regardless of their meaning. For example, in 弔 the lines that curve around the vertical line () have been replaced with bow 弓, ignoring its meaning. See also empty component.