of those categories have never been fully understood and so here we will not follow the Shuowen categories completely.
Like other writing systems, the system for Chinese evolved originally from the pictorial representation of concrete objects, so it seems logical to start here with 1) pictographs. With this category, a written representation of a horse, say, was used to represent the early Chinese word for ‘horse’, and this same principle was utilized to represent numerous other words such as ‘sun’, ‘tree’, ‘bird’, mountain’, and so on.
There was, though, a limit to the usefulness of this principle. It was fine for writing simple, concrete words, but how to write more abstract words such as those for ‘above’ or ‘basis’, for example? In the oracle bone script, ‘above’ was represented by one short stroke above a longer one, while for ‘basis’ or ‘root’ a short horizontal stroke was added low down on the vertical stroke of 木 ‘tree, wood’ to give 本. In English, graphs of this category—type 2)—are generally referred to as ‘indicative symbols’ (or similar).
In some other cases, a word was conveyed by combining several pictographs into one graph, and so in English these may be termed 3) ‘semantic compounds’. Examples of this category include 林 (two trees) for ‘forest’, or 日 ‘sun’ and 月 ‘moon’ combined together as 明 to represent the word for ‘bright’.
A further means employed to represent various words or morphemes was 4) the loan-graph principle, whereby a character was ‘borrowed’ for its sound value to represent in writing another word of the same (or similar) pronunciation. Thus, in oracle bone texts we find, for instance, the pictograph for ‘winnowing basket’ (written 其 in its stylized modern form) borrowed to represent another word of the same pronunciation meaning ‘probably’ or ‘will’. Once this happened, the reader in ancient times had to decide whether 其 in a particular context was to be taken as ‘winnowing basket’ or ‘probably/will’. In the same way, a character originally meaning ‘sunset’ (莫) was borrowed to write a similar-sounding grammatical function-word meaning ‘there is none, not any’. This sort of arrangement seems to have worked adequately at first, helped no doubt by the fact that OBI and also the very early bronze texts tended to be quite formulaic and repetitive in nature. However, as the number of such borrowings increased and also texts became more diverse in terms of content, help was needed to avoid the danger of texts degenerating into hopelessly complex puzzles. To combat this, gradually semantic markers (traditionally called ‘radicals’, but better is ‘determinatives’) were often added. Thus, because 其 ended up being used more to indicate probability or futurity than in the sense ‘winnowing basket’, 竹 ‘bamboo’ was added at the top to create 箕 for the latter (i.e., original) sense, a graph which could readily be understood to mean just ‘winnowing basket’, leaving 其 to stand for probability/futurity. The same process took place with 莫: to overcome the ambiguity of this graph when it had come to mean either ‘sunset’ or ‘there is none’, a second 日 ‘sun’ was added to create a new graph 暮 for ‘sunset’, leaving 莫 to be used for ‘there is none’. Graphs of the type 箕 and 暮 are referred to as 5) ‘semantic-phonetic compounds’ (or similar); these are by far the most common category of Chinese characters.
3 Word-Families and the Chinese Script
Note: this section, which relies extensively on the work of Japanese scholar Tōdō Akiyasu, involves much technical detail which many readers may not need; for such readers, the brief entry ‘Phonetic with associated sense’ (see Section 8 below) is recommended instead.
The application of the semantic-phonetic compounding principle led to a dramatic increase in the total number of different graphs over time. As indicated above, in semantic-phonetic compounds the phonetic element is the original element, and a semantic marker is a later addition. An important point to note is that often a particular element, while primarily phonetic, also carries a common thread of meaning that can be seen in several or a number of different graphs. This reflects the existence of words of related meaning and the same or similar pronunciation in early Chinese; grouped together, such words are known as ‘word-families’. To give a relatively simple example: the word written as 里 ‘village; unit of linear measure’ is analyzed by one scholar (Tōdō) as being made up of 田 ‘field’ combined with 土 ‘earth, ground’, originally representing a word meaning fields divided up according to a grid system, and then by extension ‘village’, representing a collection of nearby houses. This is the first of a number of words and their graphs collected together in a word-family having the core meaning ‘line, draw a line’. On this basis, we can think of ‘line, draw a line’ as the associated sense of 里 as opposed to its main meanings of ‘village’ and as a unit measure for distance. The distinction is an important one. Another word of the same linguistically reconstructed pronunciation is one for which Tōdō gives the original meaning ‘lines / veins which are visible in marble’, written 理, with 玉 (‘jade, precious / semi-precious stone’ in its abbreviated form without dot) added as a determinative (semantic marker). ‘Regulate, reason’ is a figurative extension for 理 based on ‘drawing a line’ (a straight line), and this in turn is seen in other members of the same word-family such as 裏, taken by Tōdō as 里 ‘lines’ combined with 衣 ‘garment’, giving the original meaning ‘striped inner cloth (i.e., lining) of garment’. Words in the same word-family do not necessarily involve the same written element as phonetic: in this same word-family as set up by Tōdō we find 肋, in which not 里 but 力serves as the phonetic, taken as ‘lines in (sides of) body’, i.e., ‘ribs’.
Sometimes the same written element serves as phonetic, but with associated senses which might at first glance appear to be different. The graph 肖 ‘resemble’ 1490 (q.v.) is part of a word-family in Chinese set up by Tōdō as meaning ‘small; scrape off’. At first, 肖 functioned as a graph representing a range of words of similar pronunciation and meanings which included ‘melt, dissolve’ and ‘scrape, pare, cut’. At a later stage, to reduce ambiguity in texts, determinatives were added. Consequently, for ‘melt, dissolve’, 水/氵 ‘water’ was added as a semantic marker or determinative, giving 消 (‘disappear’ is an extended sense from ‘melt, dissolve’), whereas for ‘scrape’ 刀/刂 ‘knife’ was added, giving 削. Thus, while we find ‘resemble’ given as the central meaning of 肖in character dictionaries, its associated meaning in 消 and 削 is ‘small, make small; scrape (and make small)’.
In most of the examples above, members of the same word-family share a common graph element as phonetic. Note, though, that the graph element that serves as phonetic within a particular word-family is not necessarily always the same; in fact, often it is different, as the example below shows:
word-family: ‘round, surround’ (TA1965:619-20)
垣 (‘fence, hedge’) original meaning: ‘encircling earthen wall’; the phonetic is 亘.
玩 original meaning: ‘hold and fondle precious stone in cupped hands’; the phonetic is 元.
Conversely, one and the same graphic element can have different associated senses (representing different word-families) in different graphs, as in the following example:
隅 (‘corner’) original meaning: ‘folds/creases in hills’ (禺 here represents word-family with core meaning ‘bend, complicated in detailed way’ [TA1965:307-11]).
愚 (‘foolish’) original meaning: ‘mind is stiff/inflexible’ (禺 here represents a word-family with core meaning ‘stiff ’ [TA1965:313-15]).
In addition to reconstruction of words in early Chinese by means of the methodology of historical linguistics, reliance is also placed on information gleaned from early dictionaries such as Shuowen jiezi (see Section 5 below). Often there is scope for variation in interpretation of what can be gained from this method also, and so scholars frequently differ in their analysis of a particular word or graph. To give just one example: while Schuessler acknowledges 里 may represent a member of a