The Rebus Principle and Phonetically-based Characters


            In the last file we looked the three kinds of meaning-based characters, i.e., the pictographs, the simple indicative characters and the compound indicative characters.  We concluded that, interesting and unique as they are, these types of characters are insufficient for representing the Chinese language.  A radical re-orientation is in order. 

We made an even stronger statement: any writing system that directly represents meanings, whether by pictures or ideographs, is unable to write down a full-fledged language.  Although most writing systems started out with pictographs and ideographs, sooner or later, they had to move away from the direct representation of meaning to the representation of sound instead.  Phonetic writing systems thus resulted. 

How do the two kinds of systems differ?  A diagram makes the difference clear:




                        /            \                                              

            Sound              Writing            








Sound            ß-------Writing


In a., writing represents meaning directly, while in b. writing represents sounds, which in turn represents meaning, i.e., it represents meaning indirectly. 

Why does one succeed while the other kind fail?  The reason is quite simple.  It has to do with mathematics.  Phonetic units, be it sound or syllables, are much smaller in number than the whole range of meanings any language has to represent.  There will not be a problem of running out of symbols.

But writing systems that started out as meaning-based were not simply abandoned and new phonetically based systems were created from scratch.  Rather, the old written symbols were retained and new uses were put to them.  The crucial step in the evolution of writing systems is the use of The Rebus Principle:


The Rebus Principle


Written symbols are borrowed to represent new words with the same sounds

regardless of what these symbols originally mean.


In plain language: if you don’t have a written symbol for a new word, use one that sounds the same regardless of what it means.  The most often used example to illustrate the Rebus Principle is the following hypothetical example:






abstract new words





Suppose we did not have written symbols for the words I, see and you, all of which are somewhat abstract and hard to represent with pictures.   Following the Rebus Principle, we could use existing pictographs for eye, sea and ewe (female sheep) to write the more abstract I, see and you.  We would be then using the pictographs strictly for their sounds regardless of what they mean. 

Although Chinese writing looks quite different from phonetic writing systems, the same change of orientation that occurred in other writing systems also happened here.  The meaning-based characters now only constitute about 18% of the total number of characters.  On the other hand, over 80% of Chinese characters are phonetically-based.  Let us look at the details of how this change in orientation occurred. 

Phonetic loans 假借jiajie ‘false loan’ are the “eye-sea-ewe” case in Chinese.  They are not new characters but new use of existing characters according to the Rebus Principle.   Here are two examples.  originally was the pictograph for “wheat” (the traditional style looks more like wheat).  There was no character for the verb “to come”, as it is difficult, if not impossible, to represent this motion with a picture.  It happens that  “to come” is pronounced the same way as “wheat”.  Therefore, the character for wheat was borrowed to represent the meaning “to come”.  Perhaps due to the greater use of the word “to come”, now the character is used only to write the meaning “to come”.  was originally a pictograph for scorpion (again, the traditional style looks more like it).  But there was no character for “ten thousand”, which is more abstract.  It turns out that “ten thousand” was pronounced the same way as scorpion, so the character for scorpion was borrowed to write “ten thousand”.  Like “wheat”, the scorpion meaning also is squeezed out. 

There was a problem with the use of phonetic loans, however.  Many words of different meanings would come to share the same written symbols.  The situation of homography thus results.  For the reader, this would not be as desirable as if there were separate symbols for separate words (we should not exaggerate the problem though.  We know that homophony exists extensively in spoken Chinese.  If writing is to faithfully record what is spoken, then homography is exactly what we should expect.  If people have little problem understand each other’s speech full of homophony, then there should not be any more problem for people to read each other’s writing full of homography.)

            Due to the problem of homography, attempts were then made to differentiate the different meanings of the same character by adding a meaning differentiator, i.e., the so-called semantic radical.  For example, the same phonetic component is used with a whole series of semantic radicals to distinguish the different meanings that the character came to represent:







Pronunciation of character

Meaning of character










Semantic-Phonetic Compounds 形声xingsheng ‘shape sound’ thus resulted.  In such compounds, one component is semantic and the other one phonetic.  The following examples represent the six possible arrangements of the semantic and phonetic components.  The phonetic component can be to the right (which is the most common type), to the left, on top, at bottom, outside or inside:


Whole character





Sound of phonetic

Sound of character














You may have noticed that in some cases there is a discrepancy between the sound of the phonetic component and that of the whole character.  In these cases, we have to assume that the pronunciations of the phonetic component and the whole characters were identical when the characters were created.  Later sound changes of either the pronunciation of the phonetic or the whole character has resulted in divergence in pronunciation.  That this is true can be seen in the better fit between the pronunciation of the whole character and the phonetic component in more conservative dialects like Cantonese.  Take the example of the character for ask.  In Mandarin, the phonetic component is read as men but the whole character is read as wen.  One has m sound and the other has the w sound.  Interestingly, in Cantonese, the whole character is still pronounced with the m consonant. 

            The semantic radical’s main function is the differentiation of homographs.  For representing the language itself, it is less important than the phonetic.  Try to write with just the semantic radical or just the phonetic component and see how well you will be understood.  Chances are you will not go very far with just the semantic part, while you will do quite well with just the phonetic.  After all, native speakers use wrong homophonous characters all the time without seriously compromising the message.  So if you want to sacrifice part of the character, be sure that it is the semantic radical that you let go.   Though of less linguistic value than the phonetic, the semantic radical does provide other kinds of information about the culture long ago. 








The first two characters reveal something of the material culture.  by having the wood radical, tells us that cups at that time were probably made of wood.  with the stone radical confirms that the bowls were indeed made of hard material akin to stone.  The association of thinness with sickness in tells of a time when thinness were considered sickly and also explains the traditional practice of saying someone has put on weight as a compliment.  How about the two characters with the female radical?  The first onemeans surname.  Why does surname is associated with female?  Could it be that at one time Chinese society, like almost all early societies, was matriarchal in nature?  The second one, which means jealousy, decidedly seems biased towards the female gender.