Well, I finally finished reading Imre Galambos’ Orthography of Early Chinese Writing: Evidence from Newly Excavated Manuscripts, Budapest Monographs in East Asian Studies, 2006. (Available here)
I enjoyed most of it. Below consists mainly of quotes that I think are important.
He observes that there are many graphical variations in Warring States unearthed manuscripts:
“Both modern and traditional scholars have observed significant graphical variations between character forms in Warring States China. However, they attributed most of the differences to either temporal (i.e. evolution of character structure in time) or spatial (i.e. graphical differences between the writing systems of various regions where the various scripts showed local characteristics) factors. It has not been commonly recognized, however, that variations also occurred within a corpus from the same general era and location … the differences were due not to external factors but to the flexibility in the writing habits of Warring States scribes and the tolerance of their readers.” (p. 1-2)
He argues that rather than there being one correct form, with all of the others being inferior variants, we would do better to regard them all as correct and forming a “aggregate character form.” (p. 2) He does, however, affirm that there were “dominant forms” of characters (141).
He astutely observes that when we look at the variants, “The most important and obvious practice was the retention of the phonetic element. The scribes could abbreviate or leave out almost any other part of the character, could introduce new components, yet they retained the phonetic component in virtually every instance. This realization reinforces the priority of spoken language (sound) over writing (visual form), a connection easily forgotten when it comes to Chinese writing.” (3) Later, he writes that the “the variability of character forms generally meant the variability of the semantic components. The phonetic component usually did not change. To be exact, it was the sound value of the phonetic component that remained the same, since the scribes sometimes substituted the phonetic component for another, homophonous or nearly homophonous, component.” (145)
Regarding the standardization of the script, he says “starting from the late 3rd century BC, there was a growing tendency to regularize the form, sound, and meaning of characters.” (3) But it wasn’t Lisi’s or Shihuangdi’s doing: “archaeological data reveals that the changes were, to perhaps an event greater extent, a result of a long historical evolution induced by political and administrative consolidation. The absence of any clear-cut border between the form and structure of characters immediately before and after the Qin reforms suggests that there was little distinction between the standards allegedly enforced by the government and the actual customs practiced by the people.” (3)
The Shiji only mentions the Qin reform once. The story generally comes from the Shuowen/Hanshu (31). “Rather than describing the Qin reform as it really happened, Han scholars recorded an idealized version of the event. Similarly, their view of the writing habits of their own times differs from the epigraphical evidence.” (32) Xu Shen’s small seal headings often differ from known Qin small seal characters: “the small seal script displayed in the Shuowen did not match the Qin small seal script … the Shuowen small seal script was not the original Qin seal script. Consequently, we have to distinguish between the Qin small script used at the end of the second century BC and a Han small seal script used around AD 100.” (36-9) “The differences between the character forms on the edict plates prove that the seal script in common use during Li Si’s time was not thoroughly consistent. Despite the reforms, most of the population still wrote characters with variable structures … All in all, it seems that Xu Shen, and his later interpreters, seriously overstated the effect of the writing reform.” (40)
An important point, one which William Boltz has also stated (The Origin and Early Development of the Chinese Writing System, p. 145), involves the actual purpose of Xu Shen’s Liushu 六書:
“I believe, however, that Xu Shen did not intend the liushu to explain the etymology of characters. Instead, his primary purpose was to provide a teaching tool for the study of the nine thousand characters students had to master in order to become a historian. Therefore, he intended the liushu to be used as a set of mnemonic principles for the acquisition of characters.” (54)
“… the “incorrect” folk etymologies in the Shuowen can be easily understood as mnemonic explanations.” (58)
Regarding the character 信, he writes “Xu’s explanation of the character etymologically was not correct, because 信 developed as a phonetic compound (xingsheng) character. However, since by his time the phonetic connection between 信 and 人 was perhaps not apparent anymore, he chose the huiyi principle as a mnemonic formula to aid the acquisition of the character. Indeed, even today Chinese language teachers use the same mnemonics to teach the character 信 to students.” (59)
“In fact, Xu never claimed that the liushu described the etymological origins of characters. With respect to the historical process of character formation, he provided the following explanation:
倉頡之初作書也， 蓋依類象形， 故謂之文。 其後形聲相益， 即謂之字。85
When Cang Jie first invented writing, he created graphic forms (xiangxing) according to categories; therefore these were called wen (patterns). After that, forms and sounds (xingsheng) mutually augmented each other; these were called zi.
In the above description, Xu mentioned two historical processes of character formation: xiangxing and xingsheng … Xu’s account does not disagree with our modern understanding of the evolution of Chinese characters embodied in the “three-principle theory.”87 According to the original theory proposed by Chen Mengjia 陳夢家, the three principles consisted of the xiangxing, jiajie, and xingsheng categories. Qiu Xigui has amended the xiangxing category to biaoyi 表意 (semantographs) to include not only pictographs but also characters created from semantic symbols.88 Xu Shen’s description of the evolution of characters differs from the three-principle theory only in not including the jiajie principle. However, an argument can be made that the jiajie principle is not a principle of character formation. Instead, it is a principle of the evolution of character usage. Because when a character is being used as a phonetic loan for another character, graphically it is still the original character. One could argue that since the form of the character has not changed, there has been no character-forming principle at work. Therefore, whether the jiajie principle should be counted as a principle of character formation, depends on the definition of the concept of character. If this concept comprises only the graphic form of the character, then the jiajie principle should not be included. On the other hand, the development of xingsheng characters included the process of phonetic borrowing. Every xingsheng character started its existence as a jiajie character, and only eventually did it acquire an extra component that distinguished it from the “mother” character. Therefore, once again, the jiajie principle could be regarded redundant. The organization of the dictionary itself also reveals the same approach on Xu Shen’s part.” (59-60)
Something I remain skeptical about is his assertion that the “original meaning and pronunciation usually does not completely vanish even when the character is utilized as a semantic or phonetic component in a new character.” In his note, he adds: “Both Karlgren and Boodberg (1937, p. 335) have argued that even when a component is used for its semantic value, it also carries a “weak” phonetic value, and vice versa. Accordingly, the component 之 in the above character form [志] would, at least partially, retain its semantic value and 心 its phonetic one.” (67)