Portmanteau characters in Chinese.

• Author: Branner, David Prager
• Position: Essay

Portmanteau here refers to an unusual type of Chinese character: a composite of two or more graphs for living words, all of which are to be read (in order) to give the meaning of the word represented by the whole character. It is something different from the conventional notion of the "ideograph" or huiyizi [TEXT NOT REPRODUCIBLE IN ASCII], the juxtaposition of graphs representing ideas or objects that contribute abstractly to the overall meaning of the word represented. I have shown elsewhere that characters are read through a process of recognition rather than decipherment, arguing that "complex pictograph" is a better description of the "motivation" (basis of character structure) of many graphs traditionally considered huiyi (Branner 2009). But the portmanteau is a different case. Its components are not abstract; understanding its structure depends on actually reading these components as connected words to form a phrase that defines or denotes the word.

This paper reviews a number of portmanteaux in current use and considers their place in Chinese grammatology. Such characters are of course part of the history of cursive Chinese and seem to have begun to be discussed rather late in the received history of Chinese writing, around the sixth century C.E., it is doubtful that they could be strictly the same as the huiyi mentioned in the first-century Shu[bar.o]wenjezi [TEXT NOT REPRODUCIBLE IN ASCII].

In terms of their structure and their relationship to oral words, portmanteaux embody a conception different from most mainstream characters. Their construction is more self-conscious than other character-types, which suggests that they are a later development. Their relationship to oral words is tenuous and tends to change frequently.

GRAMMATOLOGICAL MOTIVATION

The portmanteau most widely seen today is [TEXT NOT REPRODUCIBLE IN ASCII]", which represents the colloquial modal auxiliary bing 'not to need to', a contraction of buyong [TEXT NOT REPRODUCIBLE IN ASCII]. [TEXT NOT REPRODUCIBLE IN ASCII] is a portmanteau not because it represents a contraction but because it is constructed of the characters for the phrase that defines it: buyong; [TEXT NOT REPRODUCIBLE IN ASCII]. Now, the composite structure of the graph and the fact that the word it represents is a modern contraction are both well known. (1) What is less well known today is that the graph [TEXT NOT REPRODUCIBLE IN ASCII] is associated with at least two older readings that have nothing to do with the sound bang, even though their meanings are related to the decomposition [TEXT NOT REPRODUCIBLE IN ASCII] not to need'. The tenth-century Longk[a.bar]n shoujing [TEXT NOT REPRODUCIBLE IN ASCII] reads [TEXT NOT REPRODUCIBLE IN ASCII] as [TEXT NOT REPRODUCIBLE IN ASCII], our qi to discard'; the sixth-century Yanshi jiaxun [TEXT NOT REPRODUCIBLE IN ASCII] says [TEXT NOT REPRODUCIBLE IN ASCII] represents [TEXT NOT REPRODUCIBLE IN ASCII], our ba 'to stop' ~'to resign'. (2)

This diversity of [TEXT NOT REPRODUCIBLE IN ASCII]'s readings points up the unusual motivation of portmanteaux. Most Chinese characters are of the familiar "phonogram" (xingsh[bar.e]ng [TEXT NOT REPRODUCIBLE IN ASCII]) structure, combining one phonetic and semantic token each. The reading beng for [TEXT NOT REPRODUCIBLE IN ASCII] can be called phonologically motivated because it is a contraction of precisely the same two words whose graphs make up the portmanteau. But in the case of the other two readings, we are within the realm of Saussure's arbitraire du signe--only convention links qi or bit to the structure of And these other two readings are more typical of Chinese portmanteaux generally than bong is.

What is the motivation of portmanteau graphs? To answer this, it is illustrative to compare them with a different kind of playfully conceived character: ligatures or single-graph render-ings of multi-syllable words. Several of the latter were described in the 1920s by Chen Bodd (1904-1989) he says he has "collected" them from living usage (Chen 1927: l67): (3)

[TEXT NOT REPRODUCIBLE IN ASCII] [read chiling 'rescript'; used on healing talismans]" [TEXT NOT REPRODUCIBLE IN ASCII] [read hetong 'contract'; used on deeds]" [TEXT NOT REPRODUCIBLE IN ASCII] [read mili * centimeter'; there are many graphs of this sort, used in calculation'" (4) Chen presented these and other examples to defend the coining of the graph [TEXT NOT REPRODUCIBLE IN ASCII] by Doo Ding-U [Du Dingyou [TEXT NOT REPRODUCIBLE IN ASCII] ] (1898-1967) to write tushuguan [TEXT NOT REPRODUCIBLE IN ASCII] a 'library' (Doo 1927). This was part of the bubbling pot of ideas out of which the official Chinese character simplification movement later developed. Doo proposed not only [TEXT NOT REPRODUCIBLE IN ASCII] for [TEXT NOT REPRODUCIBLE IN ASCII] ("a savings of 13 strokes"), but also [TEXT NOT REPRODUCIBLE IN ASCII] for tushu [TEXT NOT REPRODUCIBLE IN ASCII] 'books' alone--[TEXT NOT REPRODUCIBLE IN ASCII] missing the bottom stroke--and a cursive form [TEXT NOT REPRODUCIBLE IN ASCII], [TEXT NOT REPRODUCIBLE IN ASCII] caught on in Japan and was popular in China for a while before the Communist Revolution, even making its way into the 1943 revision of Mathews's dictionary, handwritten by Y. R. Chao at the entry for tushuguan (Mathews 1943: 950, entry #6531). Recently, there has been a tradition that [TEXT NOT REPRODUCIBLE IN ASCII] is to be read Wan, a contraction of tushuguan, but that defeats the whole stated point of Doo's invention--to represent a multi-syllable word by a single, unique character, rather than to abbreviate the word into a single syllable. Technically, [TEXT NOT REPRODUCIBLE IN ASCII] is a kind of abbreviated ligature, but [TEXT NOT REPRODUCIBLE IN ASCII] is not necessarily a ligature at all. That is, when [TEXT NOT REPRODUCIBLE IN ASCII] represents benng, it. is indeed a ligature, because beng really consists of the two words buyong. But that is not the situation with the readings qi and biz for [TEXT NOT REPRODUCIBLE IN ASCII], where the constituents of a graph represent not the corresponding spoken word but only a definition of a spoken word. Chinese uses a single term for both ligatures and portmanteaux: hewen [TEXT NOT REPRODUCIBLE IN ASCII] combined graphs', but there is this essential difference between the two: the nature of a ligature is to compress two or more words (spoken and written words simultaneously) into the space of a single graph; a portmanteau, however, is a graphic ligature only, and it is not bound to specific words.

Portmanteaux are not part of the classical inventory of character structures. They seem closest to the huiyi [TEXT NOT REPRODUCIBLE IN ASCII] 'assembled meanings' ("syssemantic" or "ideographic") structure, although they lack what we usually think of as the abstractness of ideography. As many have observed, the ideograph is an attractive but evanescent notion with few clear-cut examples in real Chinese usage. (5) Portmanteaux differ from notional ideographs--and this is recognized by a number of Chinese sources in traditional times. Lacking explicit phonological motivation, they are constructed based on the meaning of the words (as expressed simultaneously in oral and written form) of which the graph is made up.

In addition to their distinctive structure, portmanteaux commonly exhibit two other features. They are often associated (at different times and places) with more than one word or sound. And they often seem to have come late to their modern readings, some of which have no corresponding syllable in medieval phonology. In the long era before standardization of an absolute reading pronunciation, words without a place in medieval phonology could have no undisputed identity in the historical continuum of written Chinese. Another feature of the portmanteaux is that the words they stand for often have competing graphs to represent them, which in the nature of things usually means phonograms. Where portmanteaux are involved, there are persistent problems linking written word to spoken word, because of the lack of phonological motivation. Not surprisingly, a graph that cannot be consistently linked to an oral word tends to lack stability in how it is read at different times.

LIVING EXAMPLES

The largest modern dictionaries contain hundreds of these forms, most no longer used. (6) Below I offer a tour of seven examples, all of special interest because they are associated with words known today in ordinary spoken Mandarin, after which I consider the question of how far back these graphs can be traced in the received tradition.

Consider another character containing the negative particle bu [TEXT NOT REPRODUCIBLE IN ASCII]: [TEXT NOT REPRODUCIBLE IN ASCII], a portmanteau of buzheng [TEXT NOT REPRODUCIBLE IN ASCII] 'not straight'. Today, we read it wai 'crooked, tilted to one side, off-center', a reading attested since the time of the Zihui. (7) The traditional way of writing wai is [TEXT NOT REPRODUCIBLE IN ASCII], a phonogram found in the Shuowen. (8) But our contemporary spoken word wai has not been associated explicitly with [TEXT NOT REPRODUCIBLE IN ASCII] until recently (in, for example, the eighteenth-century Kangxi ziclian), in the eleventh-century Guangyun [TEXT NOT REPRODUCIBLE IN ASCII]...