“Recycling Characters: The Significance of Phonetic Loan Characters and Substitutions in the Medieval Chinese Writing System”

Convener: Christoph Anderl, Ghent University

The notion of “phonetic loan characters,” which is used to translate both the Chinese terms jiǎjiè 假借 (‘borrowing’) and tōngjiǎzì 通假字 (‘interchangeable characters’), is of great relevance for both understanding the processes involved in the formation of the early Chinese script, as well as for studying the actual practice of writing in the Medieval period, especially those materials extant in manuscript form. In this panel we will question this English term, and define and approach the phenomena it refers to from a diachronic perspective, taking into account both the linguistic theories on this topic that were current among scholars between the Hàn and the Sòng dynasties, as well as evidence found in actual specimens of writing. A thorough analysis of the processes involved in phonetic borrowing and substitution will not only contribute to a better understanding of the medieval Chinese writing system and the phonological system behind it, but will also be of help for philologists, editors, and other researchers dealing with texts in which this is an important feature.

For our understanding of phonetic substitution and borrowing, it is important to explore how contemporary scholars defined and explained these concepts, and how they dealt with the various types of substitutions of Chinese graphs encountered in Chinese writing. The understanding of (phonetic) loan graphs in Medieval China has been deeply influenced by Xǔ Shèn’s 許慎reflections on the evolution of the Chinese script as described in the postface of his work Shuōwén jiězì 說文解字, where he introduced the term jiǎjiè 假借 as the sixth “category” of Chinese graphs. He uses this term mainly to explain the process of “loaning” a Chinese graph/character for a homophonous (or semi-homophonous) word which did not possess a written form yet. In his work he provides several examples, typically grammatical particles (xūzì 虛字) which derived their written forms from graphs representing a word with similar or identical pronunciation. For example, Xǔ Shèn indicates – without explicitly referring to the term – that qí 其 ‘basket’ (OC *gə; later written as jī 箕) was borrowed for the third person pronoun qí (OC *gə). In addition to semantically completely unrelated words sharing the same graph, however, Xǔ also cites some examples where the same graph is shared by words which are related to each other or derive from the same word family, e.g., zhǎng 長 (*traŋʔ) ‘grow; elder’ and cháng 長 (OC *Cə-[N]-traŋ) ‘long’; or the group líng 令 (OC *riŋ) ‘send’, lìng (OC *riŋ-s) ‘issue a command’ (eventually also written as mìng 命 OC *m-riŋ-s). As such, Xǔ Shèn’s explanation of loaning is rather complex, including phonological and semantic features (e.g., semantic extensions).

Later scholars, when discussing this phenomenon, base themselves on Xǔ Shèn’s analysis, but also add their own ideas. For example, the Southern Táng scholar Xú Kǎi 徐鍇 (920–974) in his commentary on the Shuōwén was the first lexicographer to elaborate on Xǔ Shèn’s definition of jiǎjiè. Xú Kǎi does not seem to conceptually distinguish between a phonetic loan and a semantic extension and groups them, following Xǔ Shèn, under the category of jiǎjiè. However, practically, and based on his experience of actually reading medieval manuscripts he does recognize a difference between phonetic and semantic borrowings. For instance, he gives an example from the Shān Hǎi jīng 山海經 in which jùn 俊 (MC tswinH) is used to write shùn 舜 (MC sywinH) without identifying any semantic connection linking the two graphs. Despite this, Xú Kǎi still refers to these cases as “distant borrowing” and considers them the result of the poor work of editors who lack proper understanding of the Chinese script. The later scholar Zhèng Qiáo 鄭樵 (1104-1162) discusses jiǎjiè primarily as a historical phenomenon, but at the same time, under a separate category, devotes some attention to phonetic substitutions that only make sense from the perspective of Medieval Chinese.

To what degree were these substitutions acceptable for him, and how do his views differ from those of earlier scholars?

The Medieval period was characterized by dramatic changes concerning the script, having to cope with significant changes in the phonological system of the spoken varieties of Chinese, the integration of thousands of new words of Indic origin, the appearance of countless new graphs and variants, the phenomenon of mass production of handwritten copies of texts, and – in the Late Medieval period – the appearance of the written vernacular. Although most medieval scholars were mainly concerned with phonetic borrowing and substitution as it occurred in older texts, manuscript evidence shows that phonetic substitution was an important feature of the writing system of the Medieval period. Although this highly complicated situation which made the reading of many text copies increasingly difficult was addressed by numerous scholars in their lexicographical works (most importantly, in the yīnyì 音義 and zìyàng 字樣 / zìshū 字書 genres), the phenomenon of phonetic substitution is only rarely directly dealt with, despite the compilers’ great efforts in determining the correct character readings. In this panel we will also discuss medieval lexicographic works, such as the mid-Táng Zhèngmíng yàolù 正名要錄 (extant in Stein 388), which defines several categories of substitutions, constituting one of the earliest sources consciously reflecting on actual writing practices in Medieval Chinese manuscript culture.

What motivated scribes to use a specific phonetic substitution, often one sharing the same phonophoric, but sometimes also using seemingly completely unrelated graphs? Were these “choices” based on a conventionalized repertoire of replacements? Or sometimes created ad hoc? Or were they based on historical relations between graphs? And to what extent did “orality” (i.e., the copyist “hearing” the word while he is copying it) play a role?

The study of loan characters / graphs during the medieval period is likewise of great value for the research on the evolution of vernacular words and for our understanding of how colloquial lexical items and function words gradually emerged in written forms, and what “shapes” they assumed before they were committed to more or less standardized forms. The use of different characters for the colloquialisms of the respective contemporary versions of Chinese does often not only reflect the use of graphical variations but can also inform us about the change of the phonetic value of a word in a diachronic perspective (e.g., the various forms of writing the precursors of 什麼 between the Late Táng and Sòng dynasties). In addition, phonetic loan characters and “local homophonous readings” can be an invaluable source for the reconstruction of dialect readings, such as the Northwestern dialect of the Medieval period. As such, the question to what extent and by what means graphs were manipulated in order to be capable of “expressing speech sounds” will be repeatedly addressed.