UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An analysis of the uses of the various forms of the human figure in the Shang script Fowler, Vernon Keith 1989-12-31

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


UBC_1989_A1 F68.pdf [ 17.33MB ]
JSON: 1.0098190.json
JSON-LD: 1.0098190+ld.json
RDF/XML (Pretty): 1.0098190.xml
RDF/JSON: 1.0098190+rdf.json
Turtle: 1.0098190+rdf-turtle.txt
N-Triples: 1.0098190+rdf-ntriples.txt
Original Record: 1.0098190 +original-record.json
Full Text

Full Text

A N ANALYSIS OF T H E USES OF T H E VARIOUS FORMS O F T H E H U M A N FIGURE IN T H E S H A N G SCRIPT By V E R N O N KEITH F O W L E R B.A., The Univeristy of Leeds, 1979 M . A . , The University of British Columbia, 1984  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T O F T H E REQUIREMENTS FOR T H E D E G R E E O F DOCTOR OF PHILOSOPHY  in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Asian Studies)  We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH C O L U M B I A September 1989 © Vernon Keith Fowler, 1989  In  presenting  degree freely  at  this  the  thesis  in  partial  University  of  British  available for reference and  copying  of  department publication  this or of  thesis by  this  his  for or  her  The University of British 1956 Main Mall Vancouver, Canada  V6T  DE-6(3/81>  1Y3  Studies Columbia  requirements that  study. I further agree that purposes  may  representatives.  thesis for financial  Asian  the  Columbia, I agree  scholarly  permission.  Department of  fulfilment of  gain  shall  be It not  the  be  by  understood allowed  an  advanced  Library shall  permission  granted  is  for  the that  for head  make it extensive of  my  copying  or  without my  written  ABSTRACT  A N ANALYSIS OF T H E USES O F T H E VARIOUS FORMS O F T H E H U M A N FIGURE IN T H E S H A N G SCRIPT  V E R N O N KEITH F O W L E R  The present thesis is a study of the design of the script of the Shang dynasty oracle bones of China. These are the earliest known examples of the Chinese script, and may be dated roughly to 1200-1051 B . C . The creators of the Shang script basically had two approaches to the representation of words: one was to represent the word indirectly, via the concept (i.e. draw the concept or refer to it graphically in some way), and the other was to represent the word directly (i.e. its phonetic shape-this was only possible after the first approach had been used, thus providing a source of graphs that could be used for their sound). In one type of graph, the so-called xingsheng or xi6sheng. both approaches are combined. In this thesis, I am primarily concerned with the type of graph in which either the whole graph or some part of it is designed with reference to the concept. In order to set reasonable bounds on the topic, I limit myself to an examination of graphs containing human figure elements. There are three basic human figure elements in the Shang script: ,and  \  and a small number of variations. The question I address in this  thesis is: What determines their distribution? At first glance they appear simply to indicate different postures: a standing figure seen from the side, a standing figure seen from the front, and a kneeling figure seen from the side. One can readily understand why there should be a standing figure and a kneeling figure, but why should there be two standing figures seen from different angles?  ii  Taking as my corpus all the graphs in Shima Kunio's Inkyo bokuji sorui containing the above mentioned human figure elements (approximately 850 graphs, or about one seventh of the total number of bone graphs distinguished to date), I systematically investigated all of them, in order to determine the relationship between the human figure elements in them and the concepts that they represent. I then sorted out about 200 of those graphs for which I felt I had been able to arrive at a correct analysis. Finally, I compared the factors determining the usage of the human figure elements in each graph to see if any consistency could be detected. I then categorized these uses, and sorted the graphs into these various categories. The body of the thesis is structured according to these various categories. The conclusions of the thesis are: 1. The element ^  , although as an independent graph is the modern character d |  , did not signify 'big,' but was chiefly used instead of 0  when the concept was felt  to be most easily or most naturally depicted from the front, i.e. where the involvement of both arms and/or both legs was felt to be particularly important to expressing the concept. 2.  The ^  element was used in graphs to do with (a) kneeling; (b) actions  typically performed in a kneeling position; (c) concepts in which kneeling could be used as a sign of inferiority, yielding, submission, subjection, etc. 3  ^  , the commonest form of the human figure, could be used in any graph  denoting any concept that was felt to have anything to do with human beings, and restrictions on its usage were determined by whether the other two elements were felt to be more appropriate.  iii  T A B L E OF CONTENTS  ABSTRACT  ii  ABBREVIATIONS  vi  ACKNOWLDEGEMENT  .  viii  :  EPIGRAPH  ix  INTRODUCTION  1  A. Aims of the thesis  1  1. M y approach to the Chinese script  1  2. Evolution of the hypothesis .!  7  B. Technical matters  15  1. List of elements  ....15  2. Corpus  ...17  3. Research method  19  4. Approach to phonological problems  24  Finals  25  Initials  25  \ / A.  28  i.a. Type of human  35  i. b. Specific human  55  ii. Human actions  71  Chapter  1.1.  iii. Body part  107  iv. Miscellaneous  114  v. Phonetic  119  2.  \  \ %  3.  \  =*-^  4.  \\ I M  150  5.  j I f  168  6.  )  I  \  / %Z / b.  7. 8. i.  11  124 -.131  (?)  172 :  1  7  8  191  Female  194 iv  ii.  Spoon  202  iii. Phonetic (?)  207  Summary of Chapter I  n.i.  209  A ' * . Standing  i.  •  2  1  1  215  ii. . .Bilateral/frontal  223  iii. Miscellaneous  249  f  $ I  2.  262  Summary of Chapter JJ  III. 1. i.  J, /  266  f  267  Real life  276  ii. Symbolic  299  iii. Phonetic (?)  ....324 331  i.  Signific  337  ii. Phonetic (?)  354  3.  i  357  4.  $l-k  361  5  £ / & ( ? )  ••••  367  Summary of Chapter HI  385  CONCLUSIONS  386  List of bone inscriptions cited or referred to  388  Bronze finding list  391  Index of graphs  397  List of works cited  403  v  ABBREVIATIONS EMC  Early Middle Chinese (the language of the Qieyun, 601 A.D.).  Glosses  Karlgren, Bernhard, 1964. Glosses on the Book of Odes. Reprinted from B M F E A 14 (1942), 16 (1944) and 18 (1946). 1970. Glosses on the Book of Documents. Reprinted from B M F E A 20 (1948) and 21 (1949). Goteborg: Elanders.  GSR  Karlgren, Bernhard, 1957. Grammata Serica Recensa. Reprinted 1972 from B M F E A 29 Goteborg: Elanders.  HY  Harvard-Yenching concordance series.  JGWB  Jiagu wenbian if ^ KaoguYanjiusuo  JWJ  KBTS  ^  *tffa , 1965. Compiled by the Zhongguo Kexueyuan  ^  (rev. and enl. from the original edition of Sun Haibo >[£ , 1934). Peking: Xinhua Shudian. Reprinted 1978 Hong Kong: Zhongguo Shuju. L i Xiaoding % , 1965. Jiagu wenzi jishi ? «§• % | f . Nankang (Taiwan): Academia Sinica. Shirakawa Shizuka fa h\ $  , 1962-1984. Kimbun tsushaku ^  (56 instalments). Kobe: Hakutsuru Bijutsukan & NJWB  Rong Geng %. fii , comp., Zhang Zhenlin i& 3% Jinwenbian  1$d. Peking:  Shuju. OBI  Oracle bone inscriptions.  OC  Old Chinese (time of the Shijing).  OJWB  Rong Geng ^ Chubanshe.  j j | , 1959.  Jinwenbian  ^  .  and Ma Guoquan j §  y  , rev. and enl., 1985.  \ fyj ^  JC  JL $*9.  Zhonghua  Peking:  Kexue  ShimaKunio 1967. Inkyo bokuji sorui \S% h -$ M f l . Rev. and enl. 1971 Tokyo: Kyiiko Sho'in. (This abbreviation is separated by a period from a following page and row number.) y  SW  Xu Shen ^ , 100. Shuowen jiezi Zhonghua Shuju.  SWDZ  Duan Yucai r% i , 1815. Shuowen jiezi zhu%ij Reprinted 1979 Taipei: Yiwen Yinshuguan; 1981 Shanghai: Guji Chubanshe.  vi  %  . Reprint 1963 Peking:  SWGL  Ding Fubao T ^ | 4% , comp., 1930.  Shuowen jiezi gulin H  £  1^ Shanghai: Gulin Jingshe | & M- 3fif % . Reprint 1960: Commercial Press. Tongjian  ^ Taipei:  Gu Jiegang 4 l * ff f l ' J, chief ed., 1966. Shangshu tongjian f * l ^ j j L ^ t . Reprint 1966 Taibei: Chinese Materials and Research Aids Service Center.  Published collections of oracle bone inscriptions are cited according to the abbreviations in Keightley 1978:229-231 (Bibliography A), except that I have transcribed them into pinyin. In addition, Zongtii- Chen Mengjia 1956, and Yingguo = L i Xueqin 1985. Translations of the Shijing are all taken from Karlgren 1950a. Note on tones in pinyin transcriptions: 1st tone is unmarked, 2nd tone is marked by an acute accent, 3rd tone by a circumflex, and 4th tone by a grave accent.  vii  ACKNOWLEDGEMENT First and foremost, my greatest debt of gratitude goes to Professor Takashima, who has guided my studies in Chinese palaeography since I embarked on them seven years ago. During this time he has also helped in supporting me by arranging various projects, and was also responsible for my successful application for a Monbusho scholarship, by getting Professor Ito Michiharu to agree to taking me on as his student. As for the present thesis, Professor Takashima not only suggested the topic, but also made numerous comments and suggestions on the successive drafts in the midst of a very busy teaching and research schedule. The other person to whom I am most indebted for guiding the thesis is Professor Pulleyblank, who was always ready to discuss matters of Chinese historical phonology with me, and helped me to put the phonetic symbols into the text. I am also grateful to him for taking over the role of acting supervisor while Professor Takashima was in Japan. For my main financial support, I am indebted to the Asian Studies Department at U B C , who provided me with Teaching Assistantships each year that I was here. I would also like to say a special thankyou to the Interlibrary Loan staff for their herculean labours in getting hold of rare oracle bone collections.  viii  ...to require a young scholar to be original is as inhumane as requiring him to be a humorist, or a poet.  A.R. Burn (Introduction to Selincourt's translation of Herodotus'  ix  Histories)  1  INTRODUCTION  A. AIMS O F T H E THESIS  1. M y Approach to the Chinese Script  The Chinese script may be approached in a number of different ways, depending on one's purposes.  Anthropologists might approach it as a cultural phenomenon,  the  distinctive hallmark of a unique civilization; calligraphers might approach it as an object of beauty; while a linguist would simply consider it as the conventional representation of the Chinese language, no different from any other conventional representation of any other language.  There is a popular feeling amongst Chinese people (who are naturally very  proud of their unique script), and also amongst foreign students of Chinese, that the Chinese script, due to its pictographic origins, is somehow more 'meaningful' than an alphabet, as if there were some mysterious interaction between the written characters and the spoken words, and one sometimes hears people say that the simplified characters 1  !Cf. Creel's (1937:159) remark that "We [who use an alphabet] have specialized on the representation of sounds; the Chinese have specialized on making their writing so suggestive to the eye that it immediately calls up ideas and vivid pictures, without any interposition of sounds." DeFrancis (1984:141) criticizes this attitude. Cf. also Chiang Yee (1938:1): "...in their written form Chinese characters not only serve the purpose of conveying thought but also express in a peculiar visual way the beauty of the thought." Admittedly one might expect this sort of remark from someone who is writing about calligraphy, but he also maintains (ibid:35) that "In the present-day style of writing, though the original image has in many cases been lost, there is still a vivid enough image to move the reader's feeling and stir associations with other characters. ...A character, being evolved from a picture, displays its meaning clearly through its appearance, even if it has no sound." These words by a Chinese writer express a feeling that is common both among Chinese people and Westerners who learn Chinese.  2  brought in by the present government are less meaningful than the old characters. However, anything in spoken Chinese that is intelligible to the ear of a literate Mandarin speaker, is also intelligible to his eye when written down , and in a scientific approach one 1  has to acknowledge that a page of Chinese written in the traditional script contributes no more to the meaning than the same text written in pinyin (though it does of course, as Professor Pulleyblank has pointed out to me, disambiguate homophones). In the normal course of reading, the Chinese reader is not constantly distracted from the total meaning of the sentences by mental images of what the characters orignally depicted, and even the student of etymology seldom pauses to think on such things. Since even those characters that originally depicted objects and animals have become stylized beyond recognition in the modern script, the uninitiated cannot guess what they represent. It is only after one learns the meaning of a character, that one can begin to imagine some resemblance to the object or animal it originally depicted. I do not think I need to labour this point, which has been amply commented on for many years, and no linguist would maintain that the Chinese script, in its present form, is anything other than a conventional representation of the Chinese language. Several years of familiarity with the Shang script have led me to the conclusion that that stage of the Chinese script bore exactly the same conventional relationship to the spoken language as the present stage of the script does, the differences being purely in form, not in nature. One should not allow oneself to be misled by the highly pictographic quality of the Shang script into proposing more complex relationships. The reason why I include this reminder here, is in order to avoid confusion over what I am trying to do in this thesis.  2  In this  thesis, I am not dealing with the relationship between the written sign and the spoken word, but between the written sign and the concept that the word refers to. I address the Professor Pulleyblank has pointed out to me that the intelligibility of the spoken and written languages may differ since, in any written language, there are always some elements left out, e.g. intonation. Intonation can disambiguate spoken sentences that might be ambiguous in the written form. I am actually indebted to Professor Pulleyblank here, who suggested that I make this clarification, and whose discussion of my first draft helped me to sort my ideas out. 1  2  3  question: how did the creators of the Shang script set about designing graphs to represent words? Naturally, then, it is vital to distinguish between character function and character design. It will make things clearer if I first give a diagram of the relationship between graph, word and concept (a triad which is comparable to the traditional Chinese xing f$ ym  -jz#  ,  |J.N  andyi fx, .respectively).  CONCEPT design / 1/ GRAPH <  \.function design  \ —WORD  function  The creators of the Shang script basically had two approaches to the representation of words: one was to represent the word indirectly, via the concept (i.e. draw the concept or refer to it graphically in some way), and the other was to represent the word directly (i.e. its phonetic shape-this was only possible after the first approach had been used, thus providing a source of graphs that could be used for their sound). In one type of graph, the so-called xingsheng " /  or xtesheng **© % , both approaches are combined. In  this thesis, I am primarily concerned with the type of graph in which either the whole graph or some part of it is designed with reference to the concept. The relationships that the above diagram summarizes may be spelt out, and related to the SW analysis of character types, as follows (using the symbol > to mean 'is represented by'):  1  Design  SW Character Type  CONCEPT > GRAPH  xiangxfng | c ^  • huiyi ^  ^  ,  should perhaps point out here that the SW classification is a later analysis, not something that we know was in the minds of the original creators of the script. I am grateful to Professor Pulleyblank for reminding me to clarify this point. 1 t  4  WORD > GRAPH CONCEPT, GRAPH WORD  xningsheng  f£  . zhuanzhu %% -k  Function G R A P H > W O R D > CONCEPT  It might also be helpful here to explain the six SW character types:  1. Xiangxing: pictographs. This method of representing words is only feasible if they refer to physical things having a definite shape.  2. Huiyi: the combining of concrete elements to suggest another concept, usually an abstract one. This was particularly used for verbs where, for example, the depiction of a person along with the object involved in the action suggests the action itself. By the time of the seal script, many elements that had been connected in the Shang script became separated, so the category of huiyi can only be clearly defined for the later script. In the Shang script, the borderline between huiyi and xiangxing is very hazy. For example, is ji / ^  e  a man pushing a plough along (xiangxfngL or a combination of 'man' +  'plough' (huiyi)? Does the evolution of the bone graph ^  , showing a person with the  foot depicted in detail, into the modern character cji jt^ , in which the person and the foot are separated, involve a change of category from xiangxing to huiyi? As far as the bone script is concerned, I do not think it is meaningful to attempt to make a distinction between these two categories.  5  3. Zhishi: the iconic representation of abstract concepts. Very few words could be so represented. Examples are shang  / _k- 'on top,' a short line on top of a long  upward-facing curved line, and x]|  / ~ F 'beneath,' a short line underneath a longer  downward-facing curved line.  From the point of view of the relationship between graph, concept and word, there is no need to make a distinction between the above three types.  4. Jiajie: rebus. This is particularly common in the early stage of the script. In order to write down a word denoting a concept (usually, but not necessarily, abstract), a graph denoting some concrete object, the word for which was pronounced the same as or similarly to the target word, was used. A common example in the bone script would be the use of H  , a stylized form of ding  *tan? 'cauldron' for the verb zhen j |  *tra n 'to test through divination.' As the script lost its pictorial quality, any type of character came to be used as a phonetic loan, whether it was originally a simple pictograph or not, so the term 'rebus' then ceases to be accurate, and one simply has to call them phonetic loans.  5. Xingsheng: phonetic compounds. This is the method that became the most popular for creating new characters, not only for words denoting abstract concepts but also for words denoting concrete concepts. It consists in the combination of an element suggesting the general area of meaning with another element the same or similar in pronunciation to the target word. Thus for example there are many women's names in the oracle bones which consist of the woman element nii pronunciation, such as jiang ^  , plus another element indicating the  I J | - , which has yang - j - 'sheep' on top as phonetic.  In the Zhou period, many characters that were originally used as phonetic loans came to be  6  disambiguated by the addition of signifies (e.g. yl added when used for shi ^  had i i  jj  'strength' signific  'force'-see GSR 330a).  6. Zhuanzhu. In his postface to SW, X u Shen gave as examples of this type of character kao "^5~ and lao.  . There has been much ink spilt on what exactly Xu Shen meant by  this category (a thorough summary of past scholarship may be found in Serruys 1957). In the body of his work, X u Shen analyses lao_ as huiyi (the bone graph is simply a pictograph of an old man-see 1.7.33) and kao as consisting of laa signific plus kao "5"  phonetic  (which, as far as the seal form is concerned, is a correct analysis, but see 1.7.33 for a discussion of earlier forms). If the term zhuanzhu is intended to describe the relationship of kao ~AT to lao  , then, purely from the point of view of graph structure (that is to  say, ignoring any etymological relationship there might be between the words represented by these characters), this is no different from xfngsheng.  By making explicit the relationship between graph, word and concept, it is thus possible to clarify the nature of the traditional SW categorization of Chinese characters. In my thesis, I have tried to be very strict about the use of these three terms. In general, I use the word 'graph' to refer to bone and bronze graphs, and the word 'character' to refer to modern characters. Since the early graphs are less standardized than the modern characters, I also find it useful to use the terms 'grapheme' and 'allograph,' modelled on the analogy of 'phoneme' and 'allophone,' when discussing the issue of graphic variants, an issue which is quite a problem as it is often hard to decide whether similar graphs are independent graphs or allographs of the same grapheme.  Whenever conducting any line of research, it is of course essential to have hypotheses and then set out to test them, modifying or discarding them as the evidence dictates. However, it is easy to pay lip-service to this oft-quoted ideal, and there is always  7  the danger of becoming too attached to a hypothesis and consequently failing to recognize counter-evidence, or dismissing it as invalid without giving it full consideration. As Pulleyblank (1985:304) has remarked: "In the long run, however, it is a losing game. If one does not actively seek for the counterevidence to one's own theories, one can be sure that others will." I think it would be useful therefore to say something about the stages my hypothesis went through before reaching the final form as presented in this thesis.  2. Evolution of the Hypothesis  The idea of graphic analysis (as understood in the present thesis) was developed and presented by Professor Takashima in his seminars on Chinese palaeography in an informal way, and he suggested to me that I might attempt to put this theory on a firmer foundation, either upholding, modifying or disproving it. The idea was, basically, that the oracle bone graphs could be profitably analysed in greater detail than had hitherto been done, and that the graphic elements separated out from a graph by this analysis could be related to different aspects of the concept that the graph represented. The important thing was to strike a balance between under-analysis and over-analysis: if one under-analyses the graphs, then one is missing out on an advance in knowledge, whereas if one overanalyses them, then one is leaving the realm of scholarship and entering the realm of imagination. For example, Professor Takashima suggested that the graph m a ^ could be broken down into the three components ^)  , ^ ^ , and \  /  ^  , and that one  could then set about determining, by comparison with other graphs in which these elements occur, whether they had any constant, inherent significance. Since these elements do in fact recur as components in other graphs, it is an interesting and worthwhile pursuit to study them and see how the hypothesis fares. On the other hand, if one were to break the  8  graphs down into the very simplest strokes, such as —  or  I  , this would probably be  going too far. Since no one had yet attempted such a thing in writing, there was no previous scholarship for me to appeal to, so this was indeed quite a challenge. On the other hand, I found it refreshing and exciting to have something new to do, where I felt I had the chance to make a significant contribution to the study of Chinese characters. To examine all the elements that occur in the Shang script would be a major undertaking. For the purposes of the present thesis, I examine the various human figure elements that occur. The reason for this choice is partly to give a unified topic to the thesis, but also because they are among the commonest elements in the script, and thus offer a rich testing ground for applying the present theory of graphic analysis. Writing is a human institution, and tends mainly to concern human affairs, so it is not surprising that a pictographic script should exhibit a large number of human figure elements. The use of the human figure in such a script thus offers a particularly rewarding field for enquiry. There are three basic human figure elements in the Shang script, \  , /fc  , and  h  , and a small number of variations. If there were only one human figure element, say  \  , then one would simply conclude that it indicated anything to do with humans, just as  there is only one tree element  % which indicates anything to do with trees or wood, the  questions I asked myself then were: 1) What determines their distribution? 2) Can their distribution tell us anything about how the creators of the Shang script approached the problem of how to create written symbols for spoken words? At a first glance, they appear simply to indicate different postures: a standing figure seen from the side, a standing figure seen from the front, and a kneeling figure seen from the side. One can readily understand why there should be a standing figure and a kneeling figure, but why should there be two standing figures seen from different angles? My original hypothesis had been to look for a symbolic significance, and I hypothesized that \  symbolized action (a person going somewhere),  f\  symbolized  9  stasis (a person standing still), while the kneeling figure yielding, inferiority or submission.  ^  symbolized such things as  However, after completing the first stage of my  research (that is, establishing, as far as possible, a correct analysis for all the graphs in the corpus), I was forced to modify this hypothesis considerably. In the first place, the  j  element clearly had several other uses (e.g. in graphs denoting types of people, people's names, and for indicating parts of the human body). In the second place, it soon became apparent to me that the use of the ^  element was in most cases determined, not by  whether the focus of the concept was on absence of motion, but rather it was used instead of  "}  when it was felt that the concept in question was more easily or more naturally  represented from the front. This is especially the case when both sides of the body are involved, as in yi ^  / 7f\\ (  )  'armpit, of which there is one on either side of the 1  body (see U.l.ii.175). As for the kneeling element, this occurs in many graphs denoting actions that the Shang would have performed normally in a kneeling position, such as eating, so it could hardly be said to have any ulterior significance in such graphs. Thus I modified my original thesis title "The Significance of the Various Forms of the Human Figure in the Shang Script" to the present one, in order to encompass all uses, whether symbolic or not. However, there remained certain minimal or near-minimal contrasts that led me to maintain my first hypothesis in a modified form. As already mentioned, ^  seems to be  largely used in graphs denoting concepts where the involvement of both sides of the body is felt to be important to their expression, and so it was desired to show both arms or both legs. Thus, for graphs denoting concepts to do with standing, it was felt desirable to show both legs. Three graphs that illustrate this in particular are: H ^ (II.l.i.172), y i \  I It  bells)' (II.l.i.165). In ^ ^  'to stand still' (II.l.i.167), and j i ^  I ^~ 'to stand' / H  'a stand (for  , a line is drawn beneath the feet to represent the ground; in  , the person is leaning on a stick, thus further suggesting the idea of standing still (the  head and torso are also depicted in more detail, but this does not seem to be relevant to the  10  concept); while in  , the feet are depicted in detail, which I think is probably intended  to emphasize further the idea of standing firmly (the tiger-top element is phonetic). Note also the graph bing H i  'side by side, together' (II.l.i.216), which consists of two  standing side by side. It is insightful to compare these four graphs with graphs that are similar but contain  \  instead of ^  . WithH  , one may compare trng. \  I JL  (I.l.ii.4), which  shows a person standing on top of a mound of earth (cf. til k primary form of ting  /J-  ), and is perhaps the  in the sense of 'straight, upright' (GSR 835h) or ting  'stick out, crop up (as of something growing)' (GSR 835i); with jji ^ compare crt ^  I  t  o  , one may  stand on tiptoe' (I.l.ii.14), which shows a person with the foot  emphasized; with yj ^ | , one may compare lio. fj  / "j^  'old' (1.7.33), which shows  an old person with long hair leaning on a stick (the long hair here is also relevant to the depiction of the concept); and with bing All. , one may compare both cong follow' (1.4.81), which shows one person following another, and bing  'to /  'to  combine' (1.4.84), which shows two people with lines across the legs joining them. Turning now to J7 , we see that apart from its use in graphs denoting concepts where kneeling was the usual position, it is also used in graphs denoting concepts of inferiority, yielding, submission, subjection, etc. Some insights into this usage can be gained by comparing such graphs with similar graphs containing ^ $  I Ik  . For example, fji  'subdue' (GSR 934a) (III.l.ii.335), which shows a hand at a kneeling  person's back, and 3 d ^  / t>3  (now written ^ f p  (JJI. l.ii.341) may both be compared with yi ^  ) 'to repress' (GSR 915a)  / ^JL , the basic significance of which is  perhaps 'to force (someone to do something)' (I.l.ii.100) and jf ^ 1  / s>L. 'come to,  reach' (GSR 681a)(I.l.ii.72), which shows a hand reaching a person. Zhuan which I think is the primary form of xun -f-  'yield' (GSR 433a) (III.l.ii.340), shows  iGSR 851a gives meanings such as 'to serve, servant, to toil, to work.' However, the graph does not show a person working, but a person being beaten. For further arguments, see my analysis of this graph in the body of the thesis.  11  one person kneeling behind another, and may be compared with c6ng  'to  follow' (1.4.81).  Apart from these contrasts, there are also some examples of graphemes which have allographs differing in the choice of human figure element. It is also interesting to examine these, and ask oneself how it was that more than one of the human figure elements was felt to be appropriate to the expression of the concept in question. The clearest cut case is that of the bone graph for jian 7L  , which is written both f  frequency (S. 107.1x58 and S. 107.3x76 respectively).  and ^  with comparable  The first form depicts a person  with the eye emphasized, thus showing that the eye is the focus of the concept.  Since  kneeling cannot be described as a position usually or most typically associated with the simple act of seeing, the variant with \  requires some explanation, to provide which I  turn to graphic analysis. M y suggestion is that it is intended to represent the concept of seeing in the sense of 'visiting' or 'having an audience.' In many of the inscriptions, jian f / L J  is used in this sense, e.g. O  H  | t 1  I  f  (Cuibian 1292)  Dingwei-day cracking, tested: Order L i to visit the Fang. First month. 1  ° f y D 7  T  y  2  y  t  V  :  £ * l  *  *  jf T\  >1 x  * * ri.  .  (Yibian 187)  if.  Dingsi-day cracking: He (?) will perhaps visit the Fang, and not meet with wji- weapons. ^In the Cuibian commentary (by Guo Moruo), jian fang is treated as the name of a tribe 'the Jian tribe.' However, comparison with other inscriptions (e.g. the next one I cite, in which jian is preceded by the adverb gi ^  ), suggests that it is not the name of a tribe, but a verb.  12  As you can see from these two inscriptions, both the standing and kneeling variants occur in this usage, but I think that the kneeling form was designed with this meaning in mind. It might be objected that the Fang, being an alien tribe, would probably have been regarded as inferior by the Shang, so that it would be strange to interpret the kneeling figure element here as implying the Shang showing respect to the Fang. One can perhaps get over this problem by appealing to the conventional nature of the script, which I began this thesis by discussing, i.e the form of the graph for jian  in these particular sentences has no  bearing on their interpretation, but is just a conventional sign for the word jian. The kneeling form simply captures the notion that 'visiting' or 'having an audience' are actions characterized in general by respect on the part of the visitor towards the person visited.  Another example is the graph for zhf 4^ other graphs), which is written both f|  and 3,  (both independently and as an element in (see III.3.2439). The standing form  simply shows a person wearing manacles, while the kneeling form emphasizes the subjection of the person thus bound. There is no observable difference in usage between these two forms. Given the caveat that the bone inscriptions are often hard to interpret, I would say that they both occur as both a verb 'to shackle, capture' and a noun 'shackled person, captive.'  {Yicun 577, Jinzhang 679, Qianbian 5.38.3), so there seems little doubt that it really is a variant of -J-fl- . Whereas 4+V seems seems to show a person lying in a coffin or grave (though the precise graphic interpretation is problematic).  13  In the case of the bone graph for xi $^ U  'captive,' JWJ 10.3245 gives variants *g  with all three human figure elements: \  ,  3024 and n.l.iii.3016) The form with \  simply indicates 'type of human,' while the  form with ^  ft  , %  (see III.5.823, 3020, 3021,  emphasizes the subjection of such people (also suggesting that the arms  are tied behind the back), and this is perfecdy understandable. What is not so clear is why there should be a variant with  y\" (which in fact is the only form in bronzes-see  O/NJWB 1362/1694-and is the form that the modern character comes from). There does not seem to be any reason why it should be necessary to depict both sides of the body for this concept. Perhaps it has the same significance as in fu.  'adult male ' i.e.  indicating an adult.  In general, however, graphs were standardized with a particular human figure element. I shall now examine a couple of cases of graphs, which are either definitely different graphemes, or whose identification as allographs is questionable.  1.  The graph  ^  (III.l.i.283) may be definitely identified as g u l ^  Xiaoding (JWJ 9.2903) also includes the graphs f and  ^  .  Li  (which JGWB 1112 also includes)  (which JGWB 1244 has separately as the non-character  ). However, when  we examine the contexts, we see some definite differences emerge:  ^  This is the commonest graph (occurring in at least twenty inscriptions), and appears  to refer to a person or group of persons. The occurrence after ling 2.846) suggests a living person, while the expression 4.18.3, where it is preceded by y_&  ?  'to order' (Xucun ^  (e.g. Qianbian  'temple ?') perhaps means 'dreams caused by the  many spirits'—it is hard to be sure. There is one example of  ^  7fj  'the Gui tribe' (not  in S, but cited by Shima 1958:417 from Dong Zuobin 1945:8.6b-it is actually Yibian 403). Dong Zuobin assigns this inscription to Period TV, on the basis of graph shapes.  14  If  ^  is the primary form of gui  'to kneel,' as I suggest it is, used as a  rebus for gui 'spirit,' then naturally the human figure in this graph would have to be Imeeling. (Although the E M C value of gui $/£> g w i a ' suggests that it was either in O C ge_ 'ffi-  * - a l or ztu > L *-aj rhyme, the little rhyming evidence there is suggests that it was * - s l rhyme. For my examination of the evidence, see i n . l . i i . 2 8 3 . )  in wei  ^  (I.l.i.a.292) Of the six inscriptions at S.46.1, two similarly worded inscriptions  contain the expression  ^  ~%  . They are both Period I inscriptions. I suggest that the l  graph be analysed as consisting of gui  phonetic and \  signific indicating 'type of  human.' By Period IV it had come to be written with the graph gui ^ texts we find mention of the Gui Fang $ L ^  thus merged with it. Compare how si ^  / ^ L , , and in soft  (e.g. Yijing H Y 38/63/3). The graph /  and si  /  collapsed as  variants of the same character (I.l.i.a.61).  ^  This graph occurs as the name of a person in four related inscriptions on Yibian  865, where it is divined whether or not he will catch Qiang. In a fifth inscription (Qianbian 6.19.8), it is the only legible graph on a fragment.  The inscriptions are too few to base firm conclusions on, but they suggest the following:  i. In Period I, gui written ^  ^  'spirit' was written  , while the name of the Gui Fang was  .  ii. By Period IV, the name of the Gui Fang had also come to be written ^  .  Yibian 6684 and Jiabian 3343. Shima 1958:417 cites a third example from Guoxue jikan (=Shen Jianshi 1935:insert before p.394), but he omits some of the graphs, and this is in fact Jiabian 3343, so there are still only two examples. l  15  iii. There are no grounds for identifying the graph be analysed as gui  'basket' on top of $i&  ft  as a variant of gu£ %^ . It may  . It could be a variant of yi ^  / f|  but it is impossible to say for sure.  2.  Another set of three graphs that differ only in the human figure element is /j?  (identity problematic ), Man ^  / ^  1  TO ^  and y i | j  / {£j . Since we find the expression  'the chief city Shang' in the bones, there can be no doubt that %  are different graphemes.  ^  and  seems to refer to a living person, so it could be a  person's name (see S. 10.4). Note however that in many inscriptions y i be interpreted as a person's name (e.g. S.43.3 d  ,  $j  j!£  could also  and S.43.4  0J  B. T E C H N I C A L M A T T E R S  1. List of Elements  There are three main forms of the human figure in the Shang script, and each may in turn be divided into a number if sub-forms.  The number of sub-forms may vary  according to the fineness of the distinctions one makes, but the general number is between 10 and 20. The smallness of this number when compared with the number of variations on the human figure in the Egyptian hieroglyphic script gives us an indication of its higher  ^ut perhaps a hewen of dingren "7"  >C  'soldiers and men,' as Professor Takashima has suggested.  16  degree of stylization compared with the latter. The Egyptian script makes many distinctions that one simply does not find in the Shang script. The following list gives the main forms and their sub-forms as I have determined them, in the order in which I deal with them in the body of the thesis, together with a simple description of what they depict. Their uses are discussed in the introduction to each section in the thesis.  LI  f  Standing, profile  2 \ 3  Body  i|  =  * — ^ - Recumbent figure  4  \\  More than one person  5  j  Sitting (rare)  6  Upside down figure  7  ^  Figure with hair emphasized  8  \  Spoon, use as rebus for 'female'  II. 1  ^  2  Standing, frontal ty  III. 1  Upside down  Kneeling, profile  2 ^  Kneeling with hands extended upwards  3 ^  ~ ^  Kneeling (or standing) with both arms shown  4 ^  ~  Kneeling (or standing) with hands crossed in front  5 % ~ p-'  Kneeling (or standing) with hands crossed behind  1  The reason for classifying these last three elements as basically kneeling is that their standing variants are rare. ' lr  17  I thus have fifteen different elements. variety of sub-forms;  ^  'f  is the commonest element and has the greatest  occurs in the smallest number of graphs, and has the least  number of sub-forms; while ^  is less common than ^  but commoner than j\  ,  and has a middling number of sub-forms. Thus the variety of sub-forms may be seen to be related to the frequency with which the element is used in the composition of other graphs.  2. Corpus  My corpus of graphs containing human figure elements is taken from Shima's rear index (S.592). Here he lists 424 graphs under the element, and 230 under the $ as being based upon \  \  element, 196 under the  element. In addition I include the graph nU $  A I ~&  . This makes a total of 851 graphs (approximately one seventh  of the total number of bone graphs that have been distinguished to date). The actual number of distinct graphemes is actually slightly lower than this, for the following reasons:  1. Some graphs are clearly allographs of the same grapheme (= 'character'). In the body of his work, Shima sometimes combines these allographs, and sometimes he does not. His approach may be termed 'cautious' rather than 'inconsistent,' since which graphs are allographs of the same grapheme, and which are distinct graphemes, is not something that one can make a priori decisions about: it requires research. Shima only combines graphs together as allographs of the same grapheme when their status as such is very obvious. In his front index (S.(9)-(12)), he only lists one of the allographs that he has so combined, so a more accurate grapheme count can be obtained by counting here.  18  2. A few graphs contain more than one human figure element, and are listed under each element, e.g. jfa  /* ^  (S. 117.4) contains both  \  and A  . Note also that the  graph \ \ \ (S.323.4) is accidentally listed twice in the index (S.591.4 and 7). Shima also has a few graphs that are listed twice in the body of the work (sometimes with different example inscriptions, but sometimes exactly the same).  3. Some of the graphs listed under rin element M  .  actually contain the graphically similar  Shima's stylized transcriptions obscure the difference, but  JGWB's more accurate transcriptions usually make the difference quite clear. However, in order to be certain whether a graph contains hi t  or r i n  , it has to be researched  into, and this is perhaps why Shima chose to lump them together rather than make a priori decisions, though I assume he was aware of the difference, since he distinguishes the two as independent graphs in the body of his work.  4. Some of the graphs contain a variety of other Bin-like elements. For instance, some of them contain dao  i  I'll 'knife,' e.g.  (S.133.1), %))  (S. 188.3). This seems a  ft little careless. There are also some other, unidentifiable squiggles, e.g. in f)  (S. 186.2).  The above reductions in the total of 851 are partially compensated for by the addition of 13 graphs (plus one from JGWB 4982) containing the element ^  ~ j*" ,  depicting a person with the hands tied behind the back, which Shima does not distinguish from nu ^  / ~j>C , but lists together under the latter in his index (S.592.3-7). Some of  these graphs are also probably allographs of the same grapheme, so there are actually less than 13 graphemes to be added. One may also add a few graphs which have not been found independently in the inscriptions, but only as components in other graphs. For example, ni * ^  / ft.  occurs  19  in If %  I Ht  (I.l.ii.1280) and  f* ^  (S.324.1), and although it has not been  found independently one may still do some useful research on it.  There is no point in trying to work out an exact number, since it is impossible to lay down the law about which graphs are allographs of the same grapheme and which are not. One can often make suggestions but, due to the lack of evidence, one cannot be dogmatic. In addition, the section of unidentified graphs in JGWB lists many graphs containing human figure elements that are not in S, but which are so rare, and whose contexts are so fragmentary or obscure, that they will perhaps never be deciphered, and are thus of no use to the present thesis. Even in S there are many such graphs, but until one has examined them one does not know whether they will come in useful or not. Accordingly, I have examined all of the graphs in the above described corpus, and then sifted out a representative selection for demonstrating the claims of the present thesis.  3. Research Method  1. A number was assigned to all the different graphs listed in Shima's front index, in the order there listed, from \  = 1 to ^  = 3340. Thus each of the graph headings in S  was given its own unique number. This enabled me to refer rapidly and unambiguously to any of the graphs in the concordance as I was doing my research. Dictionaries like JGWB and O/NJWB already provide their own numbering, and it would have been useful if Shima had done something like this himself. In the thesis, cross-references among graphs are given by Chapter, Part and Section, followed by this number. Graphs not included in the thesis are referred to by their location in published sources, such as S, JGWB and JWJ.  20  2. Research was conducted systematically on all the graphs contained in the above described corpus. This involved:  i. Identifying the graph in JGWB. JGWB serves as a standard list of oracle bone graphs and their variants, which are transcribed with a high degree of faithfulness. Shima's graphs tend to be standardized in his own handwriting, so they cannot serve as an accurate point of departure for analysis. By identifying the graph in JGWB, one can be sure that we know precisely which graph we are dealing with, and one can also see what all the variants are.  ii. Identifying the graph in JWJ. JWJ is the chief source for the identifications and analyses of oracle bone graphs up to 1965.  Most of the identifiable graphs are identified therein.  As the work of  identification proceeded, the remaining graphs became harder and harder to identify, so there has been little in the way of new identifications since then. I have in any case tried to limit myself to the more easily identifiable graphs.  iii. Consulting the SW analysis and definition along with the Duan Yucai commentary. Discussion of the SW analysis and definition is appended where necessary. It has become something of a tradition amongst Chinese palaeographers to begin their discussions of oracle bone graphs with the appropriate quotation from SW. M y reason for following this tradition is that SW, being the earliest etymological dictionary of Chinese characters (completed by X u Shen in 100 A.D.), serves as a useful starting point for analysis. The seal forms given in SW also represent the earliest standardization of the Chinese script, so it is useful to show how the bone and bronze forms developed into these seal forms. Just as E M C serves as a solid starting point for working one's way back to  21  O C , so the seal script serves as a solid starting point for working one's way back to bone and bronze graphs. The kaishu form of the script obscures many points of continuity that are still observable in the seal forms. I find that the SW analyses, whether etymologically correct or not, always provide very stimulating food for thought: it is easier to arrive at a better understanding through the process of discussing another person's interpretation, even where it proves to be erroneous, than to come up with an original understanding entirely from scratch. The SW definitions are also useful, since the meanings of words often shift their focus or even change completely over time, so these definitions, being closer in time to the Shang period (though still quite distant), help to guide us towards the meanings that the graphs were originally created to represent. In analysing an oracle bone graph, it is of course essential to know what the word that it represents meant at the time that the graph was created.  1  A word of warning due here is that many of the SW  definitions are tailored in order to use a word that sounds similar to the word being defined (the so-called shengxun % «"1  'paranomastic gloss'), or else in order to make sense of  Xu Shen's etymological analysis. Such definitions often give some peripheral or extended meaning of the character in question, and not its basic meaning.  2  Unfortunately the SW text is often ambiguous and hard to understand, and the short definitions, rather like the oracle bone inscriptions, provide very little context to aid one's understanding. In translating SW, therefore, I have referred constantly to the commentary by the Qing dynasty scholar Duan Yucai, which often helps to clarify the meaning, and I have summarized his comments in those cases where I feel they are important to understanding the SW text, in order to show how I arrived at my translation. These  ^ n e must here bear in mind, as Professor Takashima has reminded me, that we do not know when the script was created. It is quite possible that the meanings of some words may have changed between the time of the creation of the script and the earliest example we have of it, which is the oracle bones. Sometimes, however, the opposite is the case. Professor Takashima has pointed out to me the example of 2  bJL  Y  , which SW (3b.l9b) defines as ^  f<'] f t ^  , where b_i \  powk < *pak 'make w  pa wk < *prak 'cut, flay, peel' (GSR 1228) are morphologically related. r  w  22  summaries are introduced between the SW quotations and their translations by the letters 'SWDZ.' I have also noted other comments of his where they provide useful additional information. I have also tried to make the translations internally consistent to SW, by checking the SW definitions of problematic characters occurring in the definitions (though SW is not necessarily always internally consistent).  iv. Discussion of the graph, covering: a. Analysis of the graph, including discussion of the JWJ identification, with reference to subsequent scholarship where necessary. Page references are generally not given for scholarship included in JWJ under the graph which is under discussion, but only given if the material involved is in JWJ under a different graph. b. Notes on the usage of the graph in OBI, with inscriptions quoted in support where necessary. The reason for this is to show whether it is used in its original meaning, and if so, to help clarify what that original meaning was. This is important, because many graphs occur only as proper nouns (place and personal names), or in very obscure contexts, or even with no context at all, in which case my comments on the original meaning should be seen as correspondingly more tentative. By 'original meaning,' I do not mean the original meaning of the word that the graph represents, but rather what it meant at the time that the graph was created to represent it. This is the only relevant meaning for graphic analysis. Once again here it is important to bear in mind that we do not know when the Chinese script was created, so the oracle bone usage does not necessarily give us the meanings of words when the graphs for them were created.  3. The graphs were then grouped according to their common human figure element.  23  4. Graphs whose contexts are too fragmentary or ambiguous to be of any use were then removed. Graphs which I felt unable to give a satisfactory analysis for were also removed.  5. The factors cktermining the usage of the human figure elements in each graph were then compared to see if any consistency could be detected. A discussion of these factors was then pre-pended as an introduction to each Part of the thesis, and the discussion of what determines the use of the human figure element under each graph modified accordingly if necessary. In cases where an element was found to play more than one role, the Part was further divided into Sections illustrating each role. The thesis thus consists of three main Chapters which are divided into Parts for each of the human figure element variants, some of which are further divided into Sections illustrating the different roles of a multi-roled element. Since the use of ^  is determined 'negatively' (i.e. when ^  or ^  were not  felt to be more appropriate), there is not much one can say about the factors determining its usage. However, since the body of graphs containing this element is particularly large, it makes it easier to cope with if it is subdivided in some way. I have therefore decided to use it in order to show the general semantic fields of graphs containing human figure elements, such as types of humans, personal names, human actions, and body parts. I find that, when trying to work out the meaning of such graphs in context, they provide helpful guidelines. These semantic fields are also generally applicable to ^  is used in ^  ^  and  ^  (e.g.  , which denotes a body part-see II.l.ii.2423), though the usage in  graphs denoting persons' names is largely limited to  ^|  (this point is problematic, since  it seems that any graph can be used as a proper noun, i.e. person or place name).  6. A general introduction and the conclusions were then written.  24  One of the problems in attempting a thesis of this sort, is that there is still room for disagreement over the identification of many bone graphs. In order to establish the principles of graphic analysis, it is necessary to examine a substantial number of graphs. One could write whole articles exploring all the ins and outs of why a particular graph should be identified with a particular character, but this sort of detailed discussion would only serve to clutter up the thesis and reduce the number of graphs that could be covered. In order to keep the thesis down to a reasonable size, I have tried to select only those graphs whose identification is fairly problem-free, and I have endeavoured to make my analyses as brief as possible. However, I have also included a few unidentified graphs, in order to illustrate how graphic analysis may be used heuristically.  4. Approach to Phonological Problems  In order to analyse early Chinese graphs, it is essential to know which elements play a phonetic role or could play a phonetic role, and in order to understand early inscriptions, it is essential to know what words a graph could phonetically stand for. Some idea of how the graphs were pronounced at the time is indispensible, and the more accurate the idea the better. To some extent one can work with the Shijing rhyme categories, as Chinese scholars have done in the past and often still continue to, and some general idea of the initials as inferred from the E M C xiesheng contacts. However, vital information may be obscured by this 'formulaic' approach, and a more accurate reconstruction may open up possibilities that one would never have thought of while thinking in terms of rhyme categories and general classes of initials.  The problem of reconstruction may be  conveniendy dealt with in two parts: the finals, and the initials.  25  Finals  For reconstructing finals (including certain features which affect the final historically), I use the system presented in Pulleyblank 1977-78. Having said this, it should be pointed out that there is disagreement over which Shijing rhymes certain characters should be assigned to. This is particularly true for characters that do not occur as rhyme words in the Shijing, but there is also disagreement over characters which do occur as rhyme words in the Shijing, due to the use of hedge-rhymes (heyun and the fact that the wei  and zhi  »  'ait  ),  rhymes were already merging at that time.  Many characters only occur once or twice in rhyming position, thus failing to provide a really firm foundation for rhyme assignment, and the variety of rhyming schemes used in the Shijing also leads to disagreement over whether certain characters are intended to rhyme or not. Characters which do not occur as rhymes in the Shijing are usually assigned to the same rhyme as other characters in the same xiesheng series that do occur as rhyme words, if there are such. Otherwise one has to appeal to other early rhyming evidence (e.g. the Chuci), or the use of loan graphs in early texts. Their E M C rhymes can also be used as a clue in determining what their O C rhyme was likely to have been.  Initials  Professor Pulleyblank (1984:xvii) has said that "it is impossible at present, in my opinion, to make a complete reconstruction for Old Chinese without some radically new kind of evidence." This is particularly true in the case of the initials, due to the lack of contemporary systematic evidence. While the finals can be established on the basis of the  27  2. That the retroflex vowels that Pulleyblank reconstructs for E M C go back to O C clusters with r (Pulleyblank 1984:xvi).  These correspond to Karlgren's clusters with 1, but  Karlgren only reconstructed such clusters when there was xiesheng evidence for it.  3. That E M C 1- was *r- in O C (Pulleyblank 1977-78:185,1984:26).  4. That the source of the E M C retroflex initials was also clusters with r, e.g. *sr > s, tsr > ts (Pulleyblank 1962:127-130).  5. That velars normally palatalized before -ji- < -i- in E M C , e.g. zhi $ 0 tci (compare with j i ^ 0  kej). He also proposes that it was an * - l - that prevented this where velars  remain unpalatalized in grade IV, e.g. j i Iz.  kjit (Pulleyblank 1984:176). He also uses  the palatalization of velars, followed by fronting, to take E M C z- back to *j- (ibid: 175), and s- back to *x-, as in the cyclical sign xu. r\,  swit, which he reconstructs with intial  *x . w  6. The positing of  as one of the sources of E M C j - in cases where xiesheng series  show velar contacts (Pulleyblank 1962:105, where the symols *fi and y are used).  7. The identification of what he calls *l-type xiesheng series, where in E M C one finds the initials j , 5, ?, t\ d, but not t, tc, to' (Pulleyblank 1984:170).  26  Shijing rhymes, the xiesheng series do not provide systematic clues for the initials, only very vague and general clues. The earliest stage of the language for which the initials can be reconstructed with a high degree of certainty is that of the Qieyun (601 A.D.). It is for this reason that, when appealing to historical phonology as evidence in my thesis, I usually cite the E M C reconstructions first. The reconstruction I use for E M C is that of Pulleyblank 1984 (for a list of the initials and rhymes, see ibid:232-237). However, Pulleyblank has made many useful proposals for the reconstruction of O C initials. Those that I have appealed to in guiding the reconstructions that I use in the thesis are as follows:  1. The hypothesis that certain E M C initials which are in complementary distribution as regards Type A and Type B finals were originally the same in O C , but diverged due to the 1  different effects of these two types of finals.  The chief effects were (as taken from  Pulleyblank 1977-78:184-5):  Type A:  *g > *  Type B:  2  *t > tp, f > tp', *d > d?  *l>d  *l>?,j  *+ > t'  *i > p  *w > -yw  4  3  *n > ji  Type B syllables refers to those in Qieyun rhymes that fall wholly or partly into grade III in the rhyme tables, and Type A refers to those that do not. Pulleyblank proposes that the difference in OC was a prosodic feature, which he symbolizes by an acute accent for Type A and a grave accent for Type B. For a fuller explanation, see Pulleyblank 1977-78:184-5. Pulleyblank does not actually state this development, but it may be inferred from his reference to the complementarity of g- and f- in EMC. For this palatalization of dentals, see also Pulleyblank 1962:108. For this, see Pulleyblank 1984:165. J  2  3  4  28  Chapter I:  \  Part 1: )  In the oracle bone script,  ^  is the human figure par excellence. It occurs in  about as many graphs (424 in the present corpus) as  ^  and  ^  combined (196 +  230 = 426). It is unlikely then that such a widely used element should denote anything more specific than 'something to do with (or perceived as primarily to do with) human beings.' When we look at the meanings of the graphs containing this element as signific, we find that they do indeed cover a very broad spectrum of meaning. However, one can classify most of the graphs into a small number of semantic categories, and I think it is useful to do so, as it serves to focus one's thinking when trying to determine the meaning of oracle bone graphs which have not yet been identified, or the identifications for which proposed so far are problematical. That is to say, it can serve as a heuristic device. The categories which I propose are as follows:  i. Human: a. Type of human (e.g. child, name of tribe, anthropomorphic entity) b. Specific human (i.e. person's name)  29  ii. Human action (or action conceived of as being typically performed by humansthere is, needless to say, a lot of anthropocentrism here) iii. Body part, chiefly human (again, there is much anthropocentrism here, of which wei  is a particularly egregious example-see I.l.iii. 132)  I find that these categories serve as useful focal points. Graphs falling outside these categories may be simply classed as "Miscellaneous,' and there are also a few graphs in which  \  is phonetic. I shall give some examples of 'Miscellaneous' and 'Phonetic' as  Sections iv. and v. of the present Part. However, the main focus will be on the first three sections. As far as I can tell, \  ^  also potentially has the same broad range of usage as  , since its usage is mainly conditioned by the desire to depict both legs and/or both  arms (I say 'mainly' as there are a few graphs that do not appear to be susceptible to this explanation).  Thus we find graphs denoting human actions and human body parts  containing the  ^  element as signific. However, it is not clear to me why  have been used in graphs denoting types of human such as fu %z  should  (II. 1.iii. 159) and xi  As for people's names, there are graphs containing the 7 ^  (II-l.iii-3016).  element which are used as people's names, but since any graph could be used as a person or place name (or so it seems, e.g. 2 d $t is unidentified (e.g.  y  2  occurs only as a place name ), when the graph 1  ), it is hard to tell whether it was created specifically for the  person's name, or was created to denote some other word and is merely used as a person's name.  This place name is also written j i 4riH / * ^ (JGWB 493,S.469.1). I think it is reasonable to conclude that both refer to the same place, since both are hunting grounds, and both occur in Period IV inscriptions. See S.38.4, JWJ 10.3209. 1  2  30  I shall start off then by discussing the graph \ this part into five sections as outlined above.  itself, and then divide the rest of  31  1 ^  . ^  JGWB 986: rin  A.  S.1.1x697  JWJ 8.2607: rin  SW8a.la:  i ^  Ji  ^  , ^  t  i- ^  4  ^  i.  ^  J|  .  Si i L ^ . & ...  SWDZ8a.la: "F  A  A  «l  ft  1  t  *  f* *  t  ,  #L  &  £  'Being born upright is prized over being born sideways, and this is why the  graph depicts the arm on top and the leg underneath.' Tr: Ren  'man' is the most precious of all the life-forms in the universe. This is the  zhduwen form. It represent the arms and the legs. 2  Analysis The bone graph depicts in profile a person standing. There are two variants: and  \  \  . In the former, the head and arm are written as one stroke, while in the latter,  the head, torso and legs are written as one stroke, with the arm then added. Both variants survive in bronze script (  ,  ^  -see O/NJWB 1059/1308), but it is clearly the latter  that develops into the SW zhduwen form, and thence presumably into the modern form, though the modern form has, paradoxically, reverted to the composition of the former  rhis SW definition is evidendy based on a popular saying which I have so far discovered in five different places: Baihutong.Zhufa (see Chen L i 1977:257, Tjan 1949:456 H96), Xiaojing.Shengzhi, Hanshu.Xuan Yuan liu wang zhuan, and in an edict of Wang Mang, also in the Hanshu, translated by Wilbur (1943:453) as "In the nature of heaven and earth, man is most important." Tjan translates the Baihutong occurrence as "Of [all] creations of Heaven and Earth man is the most valuable." And Finally in an edict of Guang Wu, where Wilbur (1943:468) translates "In the nature of heaven and earth, mankind is most important," and T'ung-tsu Ch'ii (1965:191) "The human is the most important of all beings in the world." lr  2  This is supposed to refer to a style of script devised by Shi Zhou  King Xuan (827-781 B.C.), though SW (5a.2a) simply defines zim $0 read books.'  ^  , the Grand Historian of as d i shu If  ^  'to  32  form, with the head and arm written as a single stroke. I realise that people naturally think of the modern character  as depicting a person standing with his legs apart, but the  left 'leg' is in fact evolved from the arm. The left-hand radical form A  is truer to the  original. The chief point to note about this graph is its structural simplicity: two strokes, and a human being is depicted. This shows a very advanced level of stylisation, and is typical of the bone script in general. As the person is depicted in profile, only one arm and one leg appear (cf. four-legged animals are depicted in profile with two legs). This places certain restrictions on the use of this graph as a signific in the composition of other graphs. When a concept is felt to require the depiction of both arms or both legs, then the element d_& f\  is normally used instead. In such cases, dji j\  represents a human being in the same way as ren  does not signify 'big,' but simply  \  does. The r i n \  occasionally depicted with both arms (e.g. z M ^  I  element is  'captive'), but is  never  depicted with both legs. This fact corroborates my conclusion that, as a graph component, ty\  dl  is often merely a structural variant of rin.  \  . As independent graphs they are  of course totally separate, standing for different words and having their own distinct pronunciation. In the body of his work, Shima distinguishes r i n  \  I A.  from bj. \  /  'spoon,' but he does not distinguish them in his list of bone radicals. He collapses 1  both under the single radical rin  \  . In consequence, a number of the graphs listed  under this radical at S.591.1-10 actually contain hi  t.  and not r i n A _ . He also  mistakenly includes a few graphs where the element in question is actually dao \  I 71  "knife' (e.g. at S.591.5 he lists $1  identification as 1&  , even though at S. 194.1 he records its  \ ). Graphs containing dao 71 2  of course fall outside the scope  !S.5.2. The reason why only a few inscriptions are listed here is that most of them are at the back of the book in the ancestral section, S.539.2-544.4. F o r this identification, see JWJ 6.2107. 2  33  of the present thesis, but I have included a section on bi  tl  due to its rebus use as a  human graph.  Usage 1. To mean 'person,' probably male by default (otherwise nii ^  /  would  be used). 2. The proper name of an alien tribe, the Ren fang 3. In the formula yir6n  —  'the one person,' used by the Shang king to  refer to himself. In soft texts we find the expression yuyirdn  —  T the one  person,' and there are already examples of this in the bones, e.g. Jinzhang 124. classical expression guaren  A-  1  The  'the solitary person' is conceptually similar.  4. Used as a classifier for people, e.g. Qiang sjm r i n  ^-  A  'three  Qiang tribesmen.' However, as in Classical Chinese, classifiers are used quite sparsely in 2  OBI. One should also note that they follow the head noun, whereas in modern Chinese classifiers precede the head noun.  As for the status of rin  in Shang times, one may note that, on the one hand  they are often conscripted (deng ren ^  A - ), and on the other hand they are also often  sacrificed. There does not seem to be the specialised usage that Waley pointed out in the Analects? in which rin  are people who are people, as opposed to min fXy , who  are merely people, hoi polloi. Keightley (1969:252) says: "Jin had no technical sense in 4  either the Shang or Western Chou." However, one should note, as Professor Takashima  See Takashima 1984/85:235, n.8. For a study of the expression, see Hu Houxuan 1957 and 1981. For a study on quantifiers in OBI, see Takashima 1985. Waley 1938:27.  1  2  3  4  In modern times, Mao Zedong also had his own definition of who exactly constituted r6nmm &j  (See Mao Zedong 1957).  34  has pointed out to me, that the graph for min OBI.  has not been identified with certainty in  1  JWJ 12.3715 identifies the bone graph as min , distinguishing it from xu£ ^ I jt_ (JWJ 4.1131) 'to wink at' (GSR 293). The context is very limited, but he suggests that the phrase mao min 1  ^t  1  ^  refers  to human sacrifice (i.e. mao standing for M ^'j  'dismember'). Guo Moruo (ap. ibid.)  does not find min ^ in OBI, but describes the bronze form ^ (O/NJWB 1593/2022)as an eye with a blade piercing it. He thinks it originally referred to slaves who were blinded in the left eye (he suggests a relationship with mang slaves  @  'blind'), but fails to produce good evidence that min  ever referred to  35  Section i.a: Type of Human  ^  36  JGWB 1060: & S.12.3x7  JWJ 8.2785: £r  SW8b.3b:  %  ^  ,  1$  %  ML .  J>k )L,  Tr: fir ^ L J means 'child.' It is based on  %L <h&  .  f l  & k-^T  Jl_i [= positional variant of rin  . ], and  depicts the fontanelle on the child's head which has not yet closed up.  Analysis Karlgren (GSR 873a) says "The graph has been explained as a drawing of a baby with open fontanel [sic], but more probably it depicts the two tufts of a child's hair-dress." L i Xiaoding is of the same opinion, and quotes Liji.neize:  s.  *  z. % ,  ~k H .  n % 8  *  %  |  ,%i  "At the end of the third month, a day was chosen for shaving off the hair of the child, excepting certain portions,-the horn like tufts of a boy, and the circlet on the crown of a girl."  1  If this analysis of the graph is correct, then this custom must have extended back into Shang times, and it is of course possible that the Zhou borrowed their ritual from the Shang. Since the graph represents a standing figure, it does seem more likely that it represents a child rather than a baby. Compare zt  , in soft texts meaning  'child' (amongst several other uses-see GSR 964a), but probably originally intended to depict a baby. Since a baby cannot walk, it was perhaps felt unnecessary to depict the legs  iHY  12/45. Translation fromLegge 1885.111:473 (1(20).  37  properly. The meaning 'baby' is attested in the bone phrase  •Lady  Hao will have a baby' (Tieyun 127.1). On the other hand, in English one can also say 'have a child,' so it seems possible that the word zt meant 'child (including baby),' but that the graph represents the word by depicting a baby.  Usage Place name. There are no examples of it in its primary meaning.  61  JGWB 1149: si % ^  S.12.4xl  Not in JWJ  SW9b.l8a: |  ^  ^  &  ^  S W D Z 9b.43b: [Inserts after ojng. | f  :]  \V\ . K  , ^  ... g  &  *  : £  i  .  fa  . &  &  &  3%  . jt  &  %  )%  M  ^4  'Its  skin is thick and can be made into armour.' Tr: The si -^c  is like a wild ox but dark. The character is a pictograph. The head is the  same as in gin ^ form is  from  'beast' and li ^ 1  'mysterious mountain beast.'  [= positional variant of rin X-  : the old  ].  ,4v In texts, gin y*7 usually means 'bird,' but can also mean animal (cf. GSR 651j). In following my policy of trying to make my SW translations internally consistent to SW, I have translated it as 'beast.' 1  SW (14b.8a) defines it as  - » A fis»  v  'generic term for animals that run'(this definition  actually excludes birds, which fly). However, note that SW (4a.l8b) defines niao ^}  as  38  Analysis The usual bone form of si  is  ^  , and this is the only form that L i  Xiaoding recognizes (JWJ 9.3013), though JGWB recognizes both. The present graph has the same horn (note that it differs from the independent graph for horn: / $  jiao  , JGWB 573), but has a person's body substituted for the animal body. This  could well be the origin of the 'old form' given in SW. Compare the other 'half man-half 1  beast' graphs qiang  / 7 u and M  'sheep' and M. Qiang  and hji  ^  ^  / Jru > which bear the same relation to yang  ^  / J^_, 'tiger' as the present graph does to si ^  .  are both used as the names of alien tribes, and I think that this is  what the substitution of the human element for the animal body is intended to indicate. So we may have here genuine cases of 'abbreviated phonetics,' as opposed to the shengsheng A-  it  fa  %  •  in SW, many of which are spurious (that is to say, the 'sheep' and 'tiger'  elements are abbreviated to ^  and  $  ).  Usage Unfortunately the context is too fragmentary to tell, but the heuristic use of graphic analysis suggests that it is the name of an alien tribe.  ^  $rt?  &  > which clearly means 'generic term for long-tailed birds,' not 'long-tailed beasts,' so  Xu Shen himself is not consistent in his use of the word gjn ^t? . *I say 'old form' in quotation marks because the graphs that SW gives under this rubric do not form a particular style belonging to a distinct time and place, but are in fact quite a ragbag, and their pedigrees are unverifiable. For example, SW (2b .6b) gives ^.itt likely that the  f[t  as the 'old form' of er  element is an abbreviated form of s i 1^  , but it seems more  , so really V H . is the old form of  Y'J , not vice versa. In the present case, the antecedents of both ^ and are found in the bones, so they are equally 'old,' as far as one may tell. Clearly then one cannot accept the SW claim that one is older than the other. N  39  S. 14.4x804  S.19.2xl8 76a *f  . ^  JGWB 514 (all forms): g j a n g .  . |*  $j  S.19.3x7  JWJ 4.1325 (all forms): qiang %j  ct  SW 4a.l7a:  # * *  : i t  , ^  .*C^>L^  &  . vfe * f  $0  -h-  i  P*1  t  £  . >1 , j t  ^  A ^  ,^ $ ; MA ,  •  SWDZ 4a.35b: [Changes the definition to  jjL  'a type of sheep,' but qiang  refers to a type of person, not a type of sheep, so Duan is clearly wrong to tamper with the text like this, and Shirakawa (1974:615) says that the sentence does not read if-  in any of the SW editions or any books that quote it. If it referred to a type of  sheep, then the role of the 'person' element would become incomprehensible.] Tr: Qiang  refers to the Western Rong. They are sheep herders. The character  consists of rin A.  'person' and y&ng H P -  'sheep;' y £ n g if-  is also phonetic. The  characters denoting the names of the Man and Min tribes in the south have chong  ^  40  'insect' signific, the name of the D i tribe in the north has quan name of the M o tribe in the east has z M ^  'dog' signific, the  signific [exact identity unknown, but  probably originally referring to cat-like animals], and the character denoting the name of the Qiang in the west has yang ^ -  'sheep' signific. These are six types [of barbarian].  The characters denoting the Bo and the Jiaoyao in the south-west have rgn A  1  person  signific. I suppose the nature of the land must have some effect on the character of its inhabitants. Only the character denoting the Y i  ^  in the east has d l A  'big' in it.  here indicates 'person.' The Y i are by custom benevolent, and 'the benevolent are long-lived,' and that is where the country known as the Land of the Immortal 2  Gentlemen lies.  Confucius certainly had good reason to say:  "The Way makes no  progress. I want to settle among the Nine Wild Tribes of the East. I shall get upon a raft and float out to sea."  3  : the old form of qiang  is like this.  X u Shen notices that certain characters denoting barbarian (i.e. non-Chinese) peoples have insect or animal signifies, and implies that this is because their nature is something less than human.  Some others have 'person' signific, so they must be  reasonably human, and the Y i have d l "7\_  'great' signific, so they must be really decent  chaps. In fact, Confucius said he would not mind living among them. The SW 'old form' seems to be quite unconnected, and is either wrong or corrupt.  Analysis  Duan Yucai points out in his commentary, Xu Shen has in fact only mentioned four types. He  SB. suggests that 'six' should be amended to yi ^rx 'different.' However, if we include the Bo and Jiaoyao mentioned in the next sentence, that makes six, so perhaps there is something wrong with the sentence order here. Analects VI.21 (HY 11/6/23). I usually follow Waley's translation, but his rendering of shou here as 'secure' is misleading, so I have given a literal translation. X u shen has here collapsed two sayings from the Analects: HY 8/5/7 and 16/9/14. My translation is a collapse of Waley V.6 and IX.13.  2  3  41  The bone graph consists of an abbreviated form of yang rin  ^  'sheep' on top of  'person.' X u Shen maintains that the 'sheep' element is not only phonetic but  also signific, because the Qiang were sheep-herders. However, as Professor Pulleyblank has pointed out to me, by this logic one could also argue that the si 3(L and the hji  (see 115 in this Section) kept tigers.  kept wild oxen  The latter seems particularly  unlikely. Pulleyblank (1983:421) says, concerning the SW entry, that "the association with pastoralism is accidental and secondary," and explores rather the more likely k  relationship between the Qiang and the Jiang zfr with yang ^ -  clan, through the phonetic connection  .  The present grapheme has a number of variants with other elements added, the commonest being ~ \  and  . The first is a rope, tied round the Qiang's neck, and  the second is probably the primary form of shi 7%.  'stone,' and thus shows a stone tied  round the Qiang's neck. The Shang captured many Qiang in military raids, and used them both as slaves and in human sacrifice. Restraining devices of some kind were obviously necessary to prevent them from running away.  Various bronze graphs also show the  present graph with a rope round the neck:  (G7NJWB Fulu 1.56a.2/1.508)  ® 5C  (G7NJWB Fulu 1.21b. 1/1.171)  The modern variant fe> perhaps comes from forms like these, with the coming from the rope.  **  element  42  Usage Refers to members of the Qiang tribe, usually in a sacrificial context.  (93) * ^  See 1.4.93 (jing  115  JGWB 619: M j £  ^  )  S.27.1xl0  JWJ 5.1589: M  (this reference not given in S)  SW5a.l8a:  A £ .i  . j%  ...f4  : -± 5t  it -ff : ^ t t  iro Tr: The tiger is the lord of the mountain beasts. The character consists of hu_ ^ 7 [defined at SW 5a. 17a as huw6n fjP variant of rin  'tiger stripes'] and ^Lt  [= positional  ]. The legs [in the graph for] 'tiger' imitate those [in the graph for]  'human.' It is a pictograph. There are two old forms: ^  and  ffij .  Analysis The graph shows a tiger's head on top of a human body, and is to be analysed in the same way as the two previous graphs, si 1E> and qiang JGWB in recognizing it as a variant of hii $  . L i Xiaoding joins  / J^j 'tiger,' and comments that 'the  43  under part has already been simplified to r6n A forms of hu )ru  .' However, as L i notes, the bronze  are all full pictographs (see O/NJWB 631/773), so it is indeed curious  that the seal character should preserve only the form with rgn A . underneath.  Usage Name of a place and the alien tribe living there. It is significant that only the full pictograph j |  is used to refer to the animal (see the huo M  f  |  'catch tiger' inscriptions at S.225.2). This supports my hypothesis that the  /  JrL  \  element  is substituted for the tiger's body in order to indicate 'type of human.'  292  ^  JGWB1112:gui  ^  S.46.1x6  JWJ 9.2903: gut  SW: See under III.l.i.283  Analysis JGWB and JWJ both include the present graph as a variant of gu£ whose usual form is in the kneeling position: ^  . However, whereas  'ghost,' Jj  is used in  the sense of 'devilish' and perhaps even 'ghost' (the inscriptions are hard to interpret), the present graph is used as the name of an alien tribe: Gui fang $ L 7 j  (see Introduction  for details of usage). On the one hand, one could regard ^  ^  substituted for $  as  with  ^  to indicate 'type of human,' parallel to the last three cases that I have  just examined. On the other hand, I think one could analyse it as  signific with gui  44  EB S3  / ^  'basket' phonetic on top (for the phonetic details, see III.l.i.283). Although  appears indistinguishable from the independent graph for tian &  'field,' it is hard  to see how tian would be relevant either as signific or phonetic. Note that although the devil's head is homographous with tian in the bone form, in the seal form ^  it is  distinct. This could represent an attempt to correct this homography. The identity of this ffl  element as gui ^  is supported by the graph y i ^  / H  (= d_M if*;  'carry  on the head'-see II.l.ii.201), which may be interpreted as a person carrying a basket on the head. Here again, it is difficult to see how tian could be relevant either as signific or phonetic.  1328  J|L  JGWB 702: n i s  ^  S.211.3x65  JWJ 5.1903: M s j [ f  SW 5b.l4b: | <  1  ,1L  G  ,  \  Jfc  h.  ; Jt , & ; & 5-  S W D Z 5b.37a: The character mu variously written mu  -  e>  .  ^  .  U  L  or m i  [i.e. the writing with mu  in the Shijing [HY 55/223/6] and ^ j _  is a folk is written  in Liji.yueji [HY 19/25].  is a voracious beast. One source defines it as 'monkey.' It is like a  human. The character consists of xj£ while zhi  Kk  does not stand for 'mother,' but for a syllable  etymologization for a syllable whose meaning was not understood]. Nao * ^  Tr: The nao * ^  .  and sui  ^  'head' and § i  represent its hands and feet.  ^-  [representing the tail],  45  Analysis The graph clearly depicts a primate, so the use of the anthropoid element is quite natural.  The depiction of the head and tail distinguish the graph from a human  representation. However, in the inscriptions it occurs mainly as the name of a remote semimythological ancestor.  In some inscriptions he is titled Gao Zu  Ancestor.' Wang Guowei equates him with the D i K u 'f 1  is O C you \*BJ rhyme, while kji ig^ ^  ?  igr  'High  of the classics. N £ o  is in the corresponding rusheng rhyme jue  , so this is quite encouraging. The E M C initials, n- and k- respectively, seem rather  far apart, but note that njio.  is phonetic in the glottal stop initial word you ^  , so  there is some glottal contact here. The more verisimilitudinous forms are separated by L i as h6u 3jf|^  'monkey'  (JWJ 10.3113), though I fail to see the justification for this. He gives three examples: Tieyi 6.9, Houbian 2.31.9, and Yicun 886. Of these, the first clearly requires an 'animal' interpretation:  - * &| (Tieyi 6.9)  ...perhaps catch monkey.  The last example however clearly refers to the ancestor:  Tested: Seek harvest from Nao (with) nine oxen.  E.g. Yicun 645 (= Zhixu 37), Cuibian 1 and 2, Zo/igfu 24.7.  l  46  2635 5^  JGWB 996: shji  fa  S.402.3x3  JWJ 8.2649: xi 4 | r  SW 8a.7a:  J12,JZL  *h-».>JLA-.,S.4P'  -  i t ^ H H  SWDZ 8a. 18a: This character has the same sound and meaning as * i  .  and  . They  have now all been replaced by the character $H . Tr: Shu I**,  means 'stand.' It consists of r6n  'person' signific and d6u j i .  phonetic. It is read like shu  Analysis There is some dispute as to whether the present graph should be identified as xi or sjni 42  . L i Xiaoding opts for the former. However, the latter has in its favour  (1) that the Yupian gives the form I u~ , which consists of exactly the same elements as the bone graph, and (2) the element S W (5a. 15a) lists under zhu  jz.  phonetic, and defines a s H X Zhenyu claims that ^ shu ^sf  'tree.'  3-  also occurs as phonetic in sjni %»  (which  as if this were the signific, though it is clearly the 'to stand').  is a variant of  The basic phonetic is d6u  ^  .  Luo  , which he identifies as the primary form of / j * L phonetic and mu  It consists of ddu  ^  /  'tree'  signific. I think he is probably right. The words 'tree' and 'set up, establish' are perhaps etymologically related. Karlgren ( G S R 127j) gives shji ^ f f  a shangsheng reading for the  verbal meaning 'to plant, establish' (this is probably the meaning indicated by the cun 'hand' element ), and a qilsheng reading for the nominal meaning 'tree.' The 1  *As Duan Yucai (SWDZ 6a.21a) says: the hand is planting it.'  $']  i!  4" 4jL  %~ &  'citn  means to say that  47  original difference between s M  and nui  is perhaps that the former referred to a  cultivated tree while the latter referred to a naturally growing tree (note that the word mu usually refers to the material 'wood, timber' rather than a standing tree). However, although the Guangyun gives both a shdngsheng and a qusheng reading for shu %t  , it  ascribes both the nominal meaning 'tree' and the verbal meaning 'to set up' to the qusheng reading. For the shdngsheng reading it gives only the definition fushu the Dai Kan-Wa Jiten(5.Ill)  , which  glosses as tasuketateru 'to help stand up,' citing examples  from Han Y u , Bo Juyi and Song Lian, so this word may not be an ancient expression. However, in the only rhyming occurrence of s M  in the Shijing (HY 47/198/5), it is  a verb 'to plant,' and rhymes with shu i^ki , which is also a verb here 'to calculate' (cf. as a noun meaning 'number' it is read in the qusheng). Although the Jingdian shiwen does not have a gloss on shu ^jlt  in this ode (  3fH  . T-5 H  ) (Pan Chonggui  1983:1507), it does on sjul i^C , to which it gives the shdngsheng reading (Pan Chonggui  1983:1095) IfsM $ t  is in fact a tonally perfect rhyme here, then it  would have to be shdngsheng. I think what may have happened was that ghu originally had a shdngsheng reading as a verb, but this reading fell into disuse and the qusheng reading came to be used for both noun and verb. JGWB includes the present graph as a variant of the graphs  and S i  ,  which have the signifies 'woman' and 'kneeling person.' The kneeling person usually implies inferiority, so one might expect this to be the primary form of shu 'attendant.' Karlgren (GSR 127g) suggests that  represents a servant girl. The word  itself is probably derived from the meaning 'to stand' in the sense of 'to wait,' i.e. to stand in attendance. If the present graph may be identified with the word s M X then the  h  'attendant,'  element could be taken as indicating 'type of human.'  In OBI, \fy  and ^  are always used as 'disaster graphs.' L i Xiaoding (JWJ  8.2825) accepts Tang Lan's proposal that  stands for jian Ic-gT 'difficulties,' which  SW (13b.l5b) gives as the zhduwen form of jian  . Semantically this makes good  48  sense, but it is hard to reconcile the phonetic value of zhji  with this. So this is still  an open question.  Usage It refers to a person, but it is hard to know from the limited context whether it means 'servant' or is a person's name:  (Xucun 2.450) ...shu will die.  JGWB 3352 (unidentified)  JWJ 3.773: p i  SW 3a. 19b:  (I : i i Tr: m ^  A%  M  S  .  means 'one who does chores.' It consists of r i n  [which SW ibid, defines as diipu  'person' and pji  , a binome that X u Xuan explains as  'bothersome, chore-like' (cf. G S R 1211a:  'harassing, tiresome')]; pu  phonetic. £ j t (  'servant' signific. [There is no inscriptional  : the old form has chen  evidence for this form.]  iFor this identification, see 1.3.62.  is also  49  Note on Identification Sun Haibo puts the present graph in the unidentified section of JGWB, with the comment 'it used to be identified as pit identification.  ,' so evidently he does not accept this  However, Ye Yusen, Guo Moruo and L i Xiaoding all accept this  identification, which was first made by Luo Zhenyu. Guo Moruo points out that there is a similar graph on the bronze vessel % TjJ  /  ):  ^  1^  j&-  (O/NJWB Fulu 2.54b.2/2.713).  (which he refers to as The place where the gj  'basket' comes in the bone form is effaced in this bronze example, but  otherwise it consists of the same elements. Other bronze forms show a radical corruption,  ^ >Mt$ This example shows the T  / -f  flf  (O/NJWB 312/397)  'basket' distorted to a zt  'container,' and the xin  element underneath held up by two hands. The r6n  perhaps replaces the original  element  « f element. It seems likely that %,  variants of the same basic character. The seal element %•  and  are  in this character is evidently  corrupted from the bronze element  Analysis The bone graph basically shows a person holding a al ^  (= j i jL  ) 'basket'  with dots on top. Luo Zhenyu says that this represents the slave throwing rubbish out, which shows that it refers to a domestic slave used for menial chores. It is interesting to compare this with the Han tomb figurine of a domestic servant carrying a broom and dustpan (Wilbur 1943:facing p. 178). This suggests that cleaning was regarded as the task that most typified domestic servants. In Zhou texts, p i  commonly means 'carriage  driver,' but it can also refer to domestic servants in general. The Zuozhuan (HY 363/Zhao 7/Fu 1) lists the ten ranks of human society as wang  _I_  'king,' gong  I*  'duke,'  50  dafu 7 v  'great officer,' sM  'underling,' li %%.  dr  'menial,' liao ^ %  'helper.' This puts the pji 1  'simple officer,' zip. ^ 'labourer,' pji ^  'lictor,' yji _  s  'servant,' and l4i  pretty near the bottom. At any rate, it seems reasonable  to conclude that it referred to lowly domestic servants. Note also the following passage from ZhouliXiaguan.Sima xia.lipu:  Assistants-valets (li-po): lis sont charges des services (de propretd, tels que) balayage, enlevement de salet6s, arrosage, dans les cinq salles posterieures ou (se retire l'empereur).2  Rather than analyse the top element as van § better to follow Guo Moruo in treating analysing the xjn ^  / %  / %.  'flute; speech,' it seems  as representing the person's head, and  element separately. As Guo notes, this xjn ^  also found on top of ai£ 3jr 'concubine' and t6ng  ^  element is  'servant boy.' He regards it as  a chisel that was used to tattoo the foreheads of people who had been punished, a practice known as qtng <M> , and indeed this interpretation does group the characters containing V  this element as signific together into a category having a recognizable general meaning, i.e. punished persons.  Thus pjl 4lt_  must originally have referred to people who were  enslaved as a punishment. As for the tail appended to the sacral region, this is very curious. Luo Zhenyu explains it by referring to the SW (8b. la) analysis of wei  'tail,' which depicts a  person with a tail. X u Shen explains that 'the ancients sometimes ornamented themselves with tails, and the barbarians of the south-west also do this.' This suggests that the present graph could represent a person from a tail-wearing tribe that the Shang were wont to capture and enslave. It may be something of a coincidence, but the Tang encyclopaedia translations fromLegge 1861.5.2:616. translation from Biot 1851.2:233.  51  Shisitong ( same text at sections Dian 1002.3, Zhi 3162.2, and Kao 2590.3 of this work) refers to a border tribe in what is present day Yunnan called the Wei Pu  'tailed  Pu.' It says that they have tails three or four cun long, and when they want to sit down they first make a hole in the ground to accommodate them, for if they should break they will die. Tremearne (1912:104) reports a similar story from Africa, where the Yergum people said of the Gazum people that they had tails about six inches long, for which they had to make a hole when they sat down. One should also note the story of the descendants of Pan Hu  , the pet dog of Gao X i n shi  v§) ^  tKj  , who wore five-  coloured clothing with a tail-like appendage in memory of their ancestor (Houhanshu. Xinan Yi liezhuan). Wei Juxian (1960:43), who mentions the Shisitong and Houhanshu references, concludes that Pu was originally the name of a south-western tribe whom the Shang used as domestic slaves.  1  Deniker (1900:95) has remarked that "The costumes of  certain populations have given rise to the fable of men with tails," but notes that "Primitive 2  man has never had a caudal appendage since he acquired the biped attitude."  Usage The single occurrence (Houbian 2.20.10) is before the word M  }*•  'make  divination cracks,' so it would seem to be the name of a diviner. It is impossible to determine from this single, almost contextless occurrence, what period this diviner would belong to.  Professor Takashima, who has made a special study of graph typology  (Takashima 1988b), has informed me that this bone could belong to early or late (or Bin group or Chu iti  group), and is not necessarily a Dui l l .  group or Z i  group piece, so it seems that even from the form and style of the graphs it is difficult to assign this piece to a particular period.  ^ee also Guo Moruo 1930:284 and Liu Weimin 1975:54. The Pan Hu story is discussed by Liu Chungshee Hsien 1932. Cf. also the Norsu question "Is the tail of the Communist Party long or short?" meaning 'is the CCP strong or weak?' (Winnington 1959:62). For earlier tales of men with tails, see Monboddo 1774.1:257-267. 2  52  3143 r^\  JGWB 993: ping ^  S.481.1x4  JWJ 8.2627:  ping  SW8a.4b: 1$) Tr: Peng  ,  &  . Jk  A-  ,  ffl  means 'to help.' It consists of rin  phonetic. It is read like pii  as in p6iwei  M  .  it %  tit  &  .  'person' signific and p6ng 'keep someone company.'  [This reading suggests an etymological relationship, cf. such pairs as deng <f daj' 'wait,' xiang #t  tsrj'-dji  zi-aa"r)'~si fok zi-' 'resemble,' and the character n6ng  , which has the two readings nan. and naj  Analysis The graph consists of r i n \  'person' signific and p6ng 0  iA_  / $H  'cowrie-string' phonetic. The two elements are combined in such a way as to suggest that the person is carrying the cowrie-strings. This is merely a piece of calligraphic design. In bronzes, p6ng  is used as in pengyou  mr\  'friends and colleagues.'  'Friend'  is perhaps the original meaning, rather than the SW definition 'to help,' since this would make the  \  element more relevant, indicating 'type of human.' The bronze texts also  support the meaning 'friend.' In soft texts, the simple character peng phonetic loan.  is used as a  SW (4a. 18b) only has this character as an 'old form' of feng  'phoenix.'  !Cf. Pulleyblank 1962:232-233.  53  Usage Obscure.  3340  ^  JGWB 1065: xiong  £j  S.553.3-555.1  JWJ 8.2801: xjong 9<J  ^  . &  ){j  ,  XX £  SW8b.4a: &  , A  Tr: Xiong  'elder brother' means zhang -JL  positional variant of rgn  'person'] and kou  .  r% ...  'senior.'  It consists of  JU [=  'mouth.'  Analysis I have not yet come across an explanation as to why 'mouth' over 'person' means 'elder brother.' M y own speculation is that it symbolizes 'the one with the right to speak out,' which an elder brother would have over his younger brothers. This speculation is prompted by the graph dui  ^  I %j , which is perhaps the primary form of shuo  'speak, explain,' also read shui 'persuade' (see GSR 324q)!. The ba  /  x  over the mouth suggests the idea of separation, taking apart, analysis.  / A.  element  The semantic  connection between 'eight' and 'divide' in Sino-Tibetan is explored by Wolfenden 1939. The basic meaning of the word shuo * / J , Professor Pulleyblank has informed me, is  though as far as one can tell from the limited context (S.12.2), pM  %j  is not used in the meaning  'speak' in the bones. Li Xiaoding agrees with Lu Shixian that it is used in the two meanings v_u£ 'to inspect' and rui  in the sense of 'keen, valiant' (see JWJ 8.2789).  54  'release, loosen, explain,' just like jig.  and sju  , and yue  'pleased' (<  'released') is also related. However, it is not necessarily the case that a graph is designed with the central concept of the word family in mind. It could be designed on the basis of some peripheral usage which the designers of the script felt was easier to depict. The graph 7j , containing both the 'mouth' element and the 'separation' element, is perhaps an attempt to capture both the basic meaning 'to release, loosen' and the notion that it is a loosening done through the mouth, i.e. explaining. However, as Professor Pulleyblank has pointed out to me, there is a lack of early evidence for shuo &%J having the meaning 'speak,' so my suggestion must be regarded as rather speculative.  The analysis of the  present graph requires further study. JGWB and JWJ both include a kneeling variant J|  , but it is sharply  differentiated by context, being used chiefly as a sacrificial term (see S.44.1). JGWB carries the note to the this graph 'xiong /L  is used for zhu  phonetically impossible, and I would not identify this graph as xiong variant of zhu 'altar' (see III.l.i.265).  .' This is of course !£j at all, but as a  , from which it differs merely in lacking the sju  T  /  55  Section i.b: Specific Human  There are many graphs consisting of the person element plus a phonetic element which occur as the names of individuals. Many of these graphs cannot be identified with modern characters, t suspect that they were created specifically to refer to the individuals concerned, and that the person element in these graphs indicates that they refer to people, i.e. they are people's names. It is already well known that many of the graphs containing the woman element nji $  /  are the names of particular women but, as far as I know,  the parallel role of the person element in the names of men has not yet been pointed out. It seems quite logical that it should have such a parallel role. One reason why it has not been noticed may be the fact that the person element has so many different significances, as the present thesis shows. Another reason may be the fact that any graph, apparently, can be used as a proper noun, so there are many graphs used as men's names that do not contain  appropriate. For example, in the case of the first graph I deal with, ling ^ / £  , various  characters in this phonetic series have meanings such as 'transgress, ascend' (see GSR 898), for which a foot signific would be more appropriate, and in some bronze forms we find that a foot is added to the bottom of the \ sui  element of the modern form.  element, and it is this that evolves into the  56  The woman element, on the other hand, generally only indicates either 'type of woman' or 'specific woman,' so its role in the names of women is much more obvious. Since the graphs in this section are all used as people's names, I shall not comment on their usage unless they have additional uses or there is some problem in establishing their usage.  57  40  ^  JGWB 4509 (unidentified)  S.l 1.4x7  Not in JWJ  SW5b.l4a:&  , M  &  . Ik  , Ak £  ,%  ; £  &  . -  0  : £  ,>Jf  &>. Tr: Ling ^ £ means yu£  'to cross over.' It consists of sjn ^  SW 5b. 13b] and l i T N 'type of mushroom.' L i -fc ^  1  here means 'high' [i.e. sense of la.  , defined at SW 14b.lb as gao ping di \f?  defmes ling ^ £ a s c h i ' f f  [=l£-  [a walking radical-see  'high level land']. One source  ] 'slow.'  Analysis The present graph is overlooked in JWJ, but it is extremely similar to the bronze form of ling  , e.g. (as part of the character ling fJC):  f t *  ft  l l  1  (O/NJWB 1809/2316)  Professor  Takashima has suggested to me that I examine the reconstructions o f the pronunciation of ling  "5C andhi , which would be l i r j < * r a r j and l u w k < * r a k respectively. It would appear from these reconstructions that it is quite possible that the evolution of the top element of the bone graph into hi w  •3> was a phonetization. By phonetization, I mean that an element which was not originally intended as phonetic is corrupted into an element that has phonetic intent. Due to the post factum way in which they came into being, phonetizations are often less accurate than original phonetics.  58  The modern character ling ^  must have descended from a variant in which the foot was  emphasized, such as we find on the  , where it is used as a proper name l  (the graph occurs three times on this vessel, so I have selected the clearest example):  I think that this bronze graph should be analysed as consisting of ling ^  phonetic and  'foot' signific, and is the primary form for the word meaning 'cross over.' It is only the 7\  element on top of the modern character that comes form the bone graph ^  bone form also has the variant ^  . The  . The solution to this graph, I believe, lies in the  diamond shape, which is referred to in Chinese as lingxing ^£ -toy , the shape of the leaf of the water chestnut or caltrop (not to be confused with bfqi % ^  , Eleocharis  tuberosa, which is also translated as 'water chestnut'-this type is popularly called mati %  <ifi$  'horse-hoof).  The present graph may be analysed as a phonetic compound  consisting of r£n  'person' signific and v  , the primary form of ling  phonetic. Note that the ^  element on top is like sheng  •, as  I JL , which represents plants  in general, so this is probably a semantic hint. The analysis of the present graph into the three elements ^ , (}  and ^  might seem like an over-analysis to some people, but it  is precisely the validity of this level of analysis that I wish to test in the present thesis. If one were to regard the present graph as a simple pictograph, then one would have to describe it as depicting a person with a diamond-shaped head and three tufts on top. It is hard to imagine what sort of person this would refer to in real life. M y analysis of graphs consisting of the ^  element with another element added on top, suggests that the top  element is usually phonetic. This provides an initial rationale for breaking ^ \  and t% , and the possible connection with ling  rationale for the further breakdown of the $  'water chestnut' provides a  element into ^  !Not in O/NJWB. I was led to it by Karlgren (GSR 898b).  down into  and ^  .  The  ^  59  element in $ ^  may be regarded as performing the same signific role as the 'grass top'  in the modern character ling  . One may hypothesize that after the simple Shang  graph * % became forgotten, the Zhou made up the new character ling ^  to denote the  same word. The structural evolution of the script also obscured the original diamond shape, so that the original design of the graph for ling "5C was no longer apparent. Incidentally, it is interesting to note how phonetic elements are often joined to the top of the person element in such a way as to suggest that they represent the head. A similar phenomenon may be found in Egyptian, where the ideogram j i has a phonogram joined on top in certain verbs involving the notion of movement (see Gardiner 1957:51 §58).  JGWB. 1018: 46  107  (notinSW)  S.26.3x6  JWJ 8.2672: 46  (not in SW)  Analysis The graph consists of 'person' signific and you $ / la) phonetic. It occurs as the object of ling  -z 'to order' and ha. *T 'to summon.'  A 126 y S.27.3x4  JGWB 4403 (unidentified)  60  Not in JWJ  Analysis I am not certain what the components of this graph are, possibly jin ^ xiong ^  on top of  . At any rate, the person element is clear. The graph occurs after the verb ling  'to order,' so there is no doubt that it is a person's name. Professor Takashima has  suggested to me another possibility to bear in mind, which is that the present graph could be interpreted as a compound standing for notes that dingren  7  'the soldiers and men of Jin.' He  , though occasionally written separately, is normally written as  a compound graph, and the use of jjn "7 as a place name is supported by the river name Jinshui  133  {Houbian 1.25.3 and Jiabian 1152).  j$  JGWB 1021: 4& (notinSW)  S.27.4xl  JWJ 8.2671: T © (not in SW)  Analysis The graph consists of 'person' signific, and the phonetic is possibly zi % . The context is fragmentary, but it occurs after c6ng As  'to follow.' By analogy with other  inscriptions one would therefore expect it to be the name of a military leader. Professor Takashima has suggested to me that, as in the case of \*  as dingr6n  could  also be interpreted as a compound graph for 'men of Zi.' This hinges partly on the issue of whether the names occurring after c6ng  are the names of individuals or of groups, or  perhaps even both. Names like Wang Cheng Ji X  ^  and Zhi Guo ilt  , which we  61  commonly find in this position, would seem to be the names of individual military leaders, but they were presumably commanding groups of people.  657 4%  JGWB 474: m i l >|  S.107.2x12  JWJ 4.1197: m£i '@  Analysis JGWB and JWJ both include the present graph as a variant of m£i i f for which the usual form is ^  'eyebrow,'  . It seems to me that the present graph actually consists of  mei % phonetic and 'person' signific, and may be transcribed * X J . According to S, it occurs in the name Zi Mei £  y  fj  Prince Mei' (Xucun 1.1069 and Yibian 5394 ). The 1  phonetic element is joined to the top of the 'person' element in such a way as to suggest that it represents the person's head. Note that the graph ^  , which has the 'woman'  element incorporated (which S lists together with the present graph), occurs before the verb mi an i > / T O D 'give birth,' so this has to refer to a woman (Shima mis-transcribes the first example, Yibian 6481, as ^  , but this graph in fact also contains the woman element).  This graph may be transcribed "x^3 , though it is probably only coincidence that it has the same components as the character mei  , since in the bone graph the 'woman' element  was probably added to the phonetic element m6i @ in order to indicate 'name of a ;  woman,' whereas in the character mei -y/g , the 'woman' element is used to indicate 'female quality' (GSR 567d gives the meanings love, lovable; flatter, curry favour with'). l r  r h e Xucun example, though fragmentary, seems valid, but in the Yibian example the graph is actually  meng  I  (though it still seems to be a prince's name).  62  On the other hand it could be that the lady in the bones was so-called because she was lovable' or 'flattering,' or, as Professor Pulleyblank has suggested to me, it could be that this word was used as a name for women. Note also that Mei  occurs as a place name  in the Zuozhuan (HY 455/Ding 9/5 Zuo, Legge 1861.5.2:773). There is also a form with a kneeling woman ^  which seems to be the name of a  spirit (see S. 108.2).  789  JGWB 4277 (unidentified)  S.133.2x4  Not in JWJ  SW7a.8b: %  , ft %  SWDZ 7a.23a: Shen  Tr: Shen zhen ^ ^  &.  ^  .&  , ^  f  .  \  was the star of Jin ~§~ , not Shang. The Shang's star was chen  is the star of Shang. It consists of jing 00 phonetic. [As X u Xuan points out, zhgn ^  [representing stars] signific and tcin cannot be phonetic in shen 9  §i m, though it may be a phonetization.]  Analysis JWJ does not have the present graph, but it differs from the bronze form of can only in that the figure underneath is standing instead of kneeling, and the three 'mouth' elements become three circles:  63  (O/NJWB 905/1120)  A graph * * ^ has not been found in OBI. The bronze form is sometimes augmented by san -  / JE_ 'three,' as in the second example above, and it is this augmented form that  gives rise to the modem character. The three mouth-like or circular elements may represent the three stars of the Shen constellation, as Karlgren speculates (GSR 647a). It seems that the element  y  in OBI does not always represent 'mouth,' but sometimes simply an  object of any kind, as in pin y t) /  'objects' (JGWB 256). In the seal form, shen <fe>  @  contains the same seal element as xing J3L (  ) 'stars,' but the bronze forms suggest  that this is a later development (a folk etymology, if you like). The bone graph may be analysed as consisting of 'person' signific with primary form of shen  - | r 'the Shen constellation,' on top as phonetic.  , the  In all the  inscriptions at S. 133.2, it occurs as the name of a person who prepared some turtle plastrons for divination, e.g: cl  j§ : ^  Jiyou-day: (Signed:)  2385  T" ^  • ^  •  (Xubian 5.25.7 ) 1  Shen ritually prepared ten pairs (of shells for divination). 2  JU.  \ ^  JGWB 990:yi  A?  S.365.2x95 ^Mistakenly given in S as 5.25.6. 2  For this understanding of shi ^  , see Keightley 1978:16-17.  64  JWJ 8.2621  SW8a.2a:Kl\,feiyCT f#7,  ? & & ^ 4  J  SWDZ 8a.5b: Y i Yin ^  . JUL A , ^ ? .  should be inserted at the beginning of the definition. Si  ^t- is phonetic in the old form. [Although it is also in O C zhi rfe rhyme, the initials are very different.] Tr: Y i 4f  : [Yi Y i n J$  ^ ] is the wise man A-heng of the Yin dynasty, the one who  governed and brought to order the empire. The character consists of rin ym  to govern,  the old form of Y i  'person' and  contains the old form of si  ^L..  1  Analysis According to various ancient works, Y i Yin of Cheng Tang ^ Jia  j%,the  f  'Governor Y i ' was a minister  first king of the Shang dynasty (referred to in OBI as Shang  and also as Tang  ). A-heng f'jj  seems to have been his personal  name, though there are conflicting accounts on this point. In OBI, he is worshipped in the same way as the royal ancestors.  He is also referred to in OBI simply as Y i , and his  consort is referred to as Y i shuang 4?  T  ;fc  h  : %-  ty  ffe S  'Yi's consort.' For example:  •  (Houbian 2.38.6)  Dingwei-day cracking: It is that Y i that is cursing the rain (i.e. preventing it from falling?).  *  #  S  *  #  &  _  „ M  i  n  g  4 2 2 )  ^Note that si ?t- is sometimes used as a loan for a word meaning 'to regulate, be in charge o f in bronzes (Shirakawa 1984:364).  65  Perhaps seek rain from Yi's consort.  Since SW gives the legendary-historical figure as the primary meaning of this character, it may have been created especially for him. GSR (604a) gives the other meanings of yi either the rin  ft  as 'this' and 'a particle.' It is hard to relate either of these meanings to or the yjn T  component of the character, whereas it is quite easy to  relate the SW meaning to these components. Furthermore, neither of the GSR meanings is attested in the bones, so we do not even know if these words existed in the Shang language, and they could well be later loan usages in some other dialect. The 'person' element thus signifies that it is the name of a person, while the yrn ^ phonetic. Y i ft ?ji is in O C zhi $  a  element is probably  rhyme, while as for y|n  jwin', Professor  Pulleyblank has informed that, although there does not seem to be any rhyme evidence, the E M C front vowel strongly suggests that it was in O C zhen <rs rhyme, which is the nasal final rhyme that corresponds to the glide final rhyme zhi $ 9 • Thus although the E M C initials are rather different, the correspondence in the rhymes suggests that yjn ¥  is  intended to be phonetic in y i ft . Karlgren puts them in different phonetic series (GSR 604 and 1251), and says of y i  ft  that "the graph has 'man' and 'govern'," though this  analysis does not relate to any of the meanings that he gives, so I do not know what exactly he intended to suggest by this analysis. Serruys (1974:62) quotes a duizhen from Jiabian 562 in which he translates yi as if it were a copula, and Qu Wanli's kaoshi to this inscription says that yi  ft  here is the  same particle that one finds in the Shijing. I was surprised to see this since, apart from the fact that OBI already has the common copula w i i %- ( $ f £ ) , if y i  ft  were also a copula  it would be very strange that there is only this example of it. In fact, the graph here is written  , and is listed at S.25.4 as the graph that is identified in JWJ 3.1055 as you  , which is usually a place name. It differs from y i IT in that the hand is at the bottom  66  of the stick instead of at the top. This is the same difference as between yjn *  2888 $  .  If  and  JGWB 4385 (unidentified)  S.444.4x1  Not in JWJ  Analysis The graph consists of 'person' signific and a phonetic element which appears to comprise  (which does not appear to survive independently but may perhaps be  identified as the lower part of cheng $'order.'  ) with a flag on top. It occurs as the object of ling  Professor Takashima has suggested to me that I should examine the  possibility of this character is  without the ^  on top being ran -rT ~ f^T". The SW (9b. 14a) form of  . It seems possible that some of the bronze forms of the  element  could have evolved into this (see the graphs I cite under the next entry), though it is hard to find a meaning in the ran r% phonetic series (GSR 622) that jfV could depict, and ideally one would like to see etymology and graphic analysis corroborating each other. For the bronze form of ran f% , O/NJWB (1274/1580) gives such examples as:  ft  JI  J G W B (1136) gives <A as the bone form of ran =Fr , but J W J (5.1567) has this as z M 'bamboo.' One cannot decide from the context, as it appears to be a person's name (S .453.1)-.  67  2957  ^  JGWB 997: cheng  'person's name'  S.453.3xl5  JWJ 8.2635: cheng \ %  SW 8a.7b:  ,  U  k  , %.  .  SWDZ 8a. 18b: Now written Tr: Cheng  means 'to lift up.' It consists of rin  k  'person' signific and cheng  phonetic.  Analysis Although JGWB and JWJ identify the present graph as cheng pointed out that it contains nothing corresponding to the ^  , it should be  'descending hand' element,  so stricdy speaking one should transcribe it as the non-character ^ff^" - L i Xiaoding argues that it evolves from the graph  (S .454.1) which he regards as an earlier form, but it  seems curious that the hand, once lost, should reappear in the modern form. He describes the earlier graph as showing a person lifting up an object, and says that this evolved into the commoner form f ft in the same way that  evolved into bao IT /  (see JWJ  8.2611). There is an example of the earlier graph in bronzes, and also of an apparently intermediary form:  (O/NJWB 1077/1332)  68  L i confesses (JWJ 4.1408) that he does not know what object J^j.  represents. Professor  Takashima has suggested to me that this graph, which does not survive as an independent character in SW, depicts a balance.- It could be the primary form of cheng  in the sense  of 'a balance.' If so, one could say that cheng \ ^ may have evolved from ^  by the  replacement of the simple phonetic element *  . The  with the compound element  present graph may be analysed as consisting of 'person' signific and  phonetic  (whatever the latter's identity). It is used in OBI as a person's name, so there are no cogent reasons for identifying it with a modern character at all. The graph  on the other hand  is used as a verb (or nominalized verb) in the expression ruii cheng ^2 estimated' occuring in prognostications.  1  'as (the king)  The graph shows a person lifting up a balance  (so the primary meaning was perhaps 'to lift up>to weigh>to estimate'), and the balance also serves as phonetic. The balance Jf^ is very rare as an independent graph in OBI (S.453.3 has three examples ). Mostly one finds the augmented form ^  (S.454.2 has  2  113 examples), in which a hand is added. This is the SW (4b. lb) character cheng  ,  which has the same pronunciation and meaning as cheng \ \ , being defined as bing n  ^p-  'to lift up together.' X u Shen analyses it as containing gou Jft-  However, ^  is $\  is clearly an independent element, and if anything one should say that g6u doubled. Furthermore it is hard to reconcile X u Shen's analysis with the  meaning of g6u J}, cheng  abbreviated.  'framework,' which seems to have no bearing on the meaning of  . Professor Takashima has suggested to me that g6u  was originally  designed to depict a counterbalance, and that the 'framework' can be considered related to it.  1  See Takashima 1984:32.  _  The Qianbian 7.1.3 example is actually written .R. and identified in JWJ (4.1403) as zjd & context is too fragmentary to see how it is being used.  2  • The  69  2959  JGWB 1019: Ht  (not in SW) 'person's name'  S.453.4x20  JWJ 8.2671: ft  (notinSW)  Analysis The graph consists of 'person' signific, with a phonetic element that consists of one or two M  'growing grain' on top of $\ / - $ • ( = cheng  'balance'?) This  phonetic element has not been found as an independent graph in OBI, but it could perhaps be the ancestor of the modern character cheng  The addition of the 'grain' element  perhaps indicates the usual commodity that was weighed. The graph occurs before the verb dl ^) / & 'to take, to bring,' so there is no doubt that it refers to a person. Thus although the graph has not been identified with a modem character, the heuristic application of graphic analysis combined with the contextual corroboration enables us to feel fairly certain that the present graph was created in order to write down a person's name, and that this is not a loan usage.  JGWB 1020: 1*  2961 7 A S.454.1x3  A* JWJ 8.2673: I#  Analysis  (notinSW)  (not in SW) 'person's name'  70  The graph consists of 'person' signific, and a phonetic element that consists of djii $  / j|  on top of ^  /  (= c h e n g ^ & ?).  This phonetic element occurs  independently, probably as a place name (S.454.1). Although it is quite clear what the components are, it is not clear how they interact. I think perhaps the dju  element is  phonetic and the cheng -rr element signific, so the meaning would presumably be something to do with balances. It could perhaps be the primary form of zhui ^ on a balance,' which according to the Guangyun is also written phonetic. Note that JGWB includes  'weight  with zhui  i|[ , i.e.iijf. , as a variant of the present graph.  The present graph occurs as the subject of an illness, so there is no doubt that it is a person's name.  (No number) $j  JGWB 656: ji | p  S.568x238  JWJ 5.1749: ji  Analysis JGWB and JWJ both include this graph as a variant of ji ^ is always used as the name of a diviner, while j $  / fp , but curiously it  is never so used. I would say that the  present graph was specially created for the name of the diviner, and is not necessarily a variant of ji $  . As Professor Takashima has pointed out to me, this strengthens the  interpretation that the graphic elements attached to ^ also that \  and \  probably served as phonetic, and  were, on the whole, strictly distinguished.  71  Section ii: Human Actions  By far the greatest number of graphs in which the rin  ^  element occurs as  signific denote actions which are typically performed by human beings.  Actions  performable by other animals are also interpreted from an anthropocentric perspective. It should be noted that it is not the rin action. The rin  \  ^  element itself that conveys the sense of  element simply denotes human participation, and it is the other  elements combined with it that serve to convey the action. This must be so, for if the rin ^  element inherently signified 'action,' then one could not explain how it could be used  as signific in many graphs that do not denote actions. The rin  \  element thus serves  only as a pointer towards the possibility that the graph in question denotes a human action. It does not in itself represent the action. For example, in jf j$ is the hand X  that signifies the action, while rin  reached, and in jian % and the rin  ^  ^  /  'to reach'(no. 72) it  just shows the person being  I $u 'to see' (no.655) it is the eye that is the focus of the action,  element merely suggests that it is a human being that is doing the seeing.  (  72  4  ^  JGWB 1034: ting £  S.5.3xl0  JWJ 8.2709: ting J_  SW 8a.l7a: 1  ,-|r  ^ 4£ £  A>  .  .  A-  , £  ; i  , ^  ^  .  ^  0  |  'earth,'  'official.'  Tr: Ting JL 'outstanding' means shan -H" shi  Jj  Rj ...  SWDZ 8a.46a: In view of the second meaning, the lower element must be Iii not shi " i  %  'official.'  Shi i  'good' It consists of ren  'person' and  means shi -jf^ 'to serve' [but X u Xuan explains that  'person' on top of 'official' indicates  'standing upright']. One source  says that it depicts something springing up out of the ground. In the Kangxi dictionary, rin  is classified under shi d r , while ting £  is  classified under tti J L , with a note to the effect that Xu Shen was mistaken in interpreting the lower element as shi " i r , and in fact the bone graph corroborates the Kangxi's and Duan's opinion.  Note that although ting  and r e n  homographous, they were still very distinct in the seal script.  are now  almost  1  Analysis The bone graph shows a person standing upright on top of a mound of earth, and is probably the primary form of ting  in the sense of 'stick out, crop up (as something  growing), straight' (definitions from GSR 835i). The first meaning given in SW, 'good,' is to be interpreted as 'outstanding,' as L i Xiaoding suggests: ^  ^jj  used to think they were identical in the modern script, but Professor Takashima has pointed out to me that there is a difference, which I had failed to notice, i.e. the relative length of the middle and lower strokes 1 t  is opposite. I note however that Li Xiaoding writes them both as seems that not everyone observes this distinction.  4c (for rin, see JWJ 14.4297), so it  73  ?|  ^ - ^ I f ^ r f ^ r ^ - ^ - ^ ^  ( i t  d e  P  i c t s  boldly, and hence by extension it comes to mean good. Mandarin usage of ting  , as in ting Mo.  L i Xiaoding notes that, while h ting  ^  1  a  P  e r s o n  )  standing out  Compare the contemporary  'outstandingly good.' /  shows a person standing from the front,  shows the same thing from the side. However, the difference in the human  figure element used is not the only difference. L i ground, whereas ting  ^  ^  shows a person standing on flat  shows a person standing on top of a mound of earth, and this is  probably essential to the idea of 'standing out.' Flat ground would fail to convey this idea. In i i  ^  , da  ^  Evidently in ting  is used to emphasize the fact that both feet are planted on the ground. ^  it was not felt necessary to emphasize this. The emphasis is not on  the standing, but on the elevation of the standing. JGWB includes a couple of examples under ting ^  in which the ground is level,  and JWJ also includes these forms, which would provide a minimal contrast between the r6n  ^  and dj|.  >^ elements, but I am not certain if they should be included. Their  limited context does not help to resolve the issue.  Usage In eight of the ten inscriptions it seems to be the name of a person (object of ling 'order'), so the inscriptions do not help us with the original meaning. In the other two inscriptions (Yizhu 524 and Yibian 5582), as Professor Takashima has pointed out to me, it seems to be a verb, though it is hard to say what it means.  14  ^  S.6.3xll  JGWB 988: ai  ^  74  JWJ 8.2615:  <fe  SW 8a.lb:  %t  &  •  , jt  SWDZ 8a.2a: Zhong JJF should be zhong  .  $  l]  : it  , meaning gen,  £  4:  ^  'heel.' [The reason  why Duan makes this quibble is that SW distinguishes between zhong  'heel' (2a.21a)  and zhong %"% 'follow in the footsteps of (2b. 16b), but they are of course simply the nominal and verbal uses of the the same word, and are now both written zhong ^ character sheng j p  .] The  'phonetic' should be edited out: JJI j£T has always been in the  sixteenth rhyme category [OC zhi ^  rhyme], while zhi  is in the first [OC zhi  rhyme]. Tr: Q |  means 'to raise the heels.' It consists of r i n  JLL phonetic.  : the old form of QI j L contains z^  'person' signific and zh£ 'foot.'  Analysis The graph depicts a person standing on tiptoe, which is the basic meaning of gi "j^" , so the SW definition 'raise the heels' accurately describes its primary meaning. It is from this that the modern meaning 'to strive' develops. The foot is not phonetic (and in fact cannot be, as Duan rightly points out), but an integral part of the whole graph. However, already in the seal form, the foot has become separated from the person. This structural change effectively obscured from X u Shen the original composition of this character. L i Xiaoding notes that the disproportionately large size of the foot shows that this is the focal point of the graph. Compare the large eye in jian ^ the big ear in wen  /  'to hear.'  The present graph may be contrasted with ju A  / %j 'to see' and  / jj|  (II.l.i.165), where the da  element with both feet drawn in suggests standing firmly (the 'tiger head' element is  75  phonetic), while the present graph conveys the idea of stretching up on tiptoe and striving. The contrast is similar to that between ting \  and! 4  .  Usage Probably an extension of the basic meaning, as in the expression rri kut ^ ^  ^  /  , which occurs several times and could be interpreted as 'raise lances' (i.e.  transitive use of (j£  ), but the exact interpretation is problematical.  21 ^  'read zhi jg,  JGWB 1501: shi %  , sense of zhi | & 'i  S.7.1x559  JWJ 12.3737: di go  SW12b.l6a: ^  , 5.  SWDZ 12b.34a: D l definition Tr: D i ^  A » . A l &  T  %  here means dl  —  •••  'to arrive.' [Duan also inserts a second  'root' from the X u Kai edition.] means zhi ]£_ 'to arrive.' [It also means 'root.'] It consists of shi  [defined at SW 12b. 15b as the Sichuan term for a precariously perched boulder] with a line attached beneath. The line represents the ground. Xu Shen's definition of shi  is amazingly recherch6, and seems to have been  devised with his explanation of dl &s in mind. He quotes the line ^  7tz R  "And their fame resounded like an avalanche" from Yang Xiong's Jiechao 2  \ H  fefy  ,  but the present text, preserved in the Honshu, has the character f & . According to Guo  JThe Guangyun defines zhi ffo t p i ' as jf jk; 'level,' translation from Knechtges 1982:52.  'to cause to arrive, to send.'  76  Moruo, sjii $j  originally depicts a spoon (see JWJ 12.3723). Whatever it depicts, I  think there can be litde doubt that sM  d?ia < * d a j '  is phonetic in d i %  t e j ' <  *t6j', so X u Shen's 'semantic compound' analysis is unnecessary anyway.  Analysis The graph shows a person carrying something in his hand.  A number of  identifications have been proposed, of which the most generally accepted is di &  ,  understood in the sense of zhi l&* , which in soft texts means 'to bring about,' but is understood in OBI as meaning 'to bring,' or as zhi  'to cause to arrive.' If we identify  the graph primarily as sju $v , as some scholars have done, then we can accept L u Shixian's proposed relationship with !£ ^  dej < * d a j  'carry in the hand.' Note that s M  , as far as one may gather, has always been homophonous with sju  , and also  depicted a spoon, according to Guo Moruo.  Usage It is most naturally translated as 'to take' or 'to lead.'  67  JGWB 589: he.  ,=M %  S. 13.4x28  JWJ 8.2629: ^  = M ffi = h£ $  JWJ 5.1823: ^  = yjn JZJ ' = dan %  SW8a.5a:  ,%  AJ  . /A L  'carry on shoulder' 'carry on shoulder'  , ^  1£  'carry on shoulder'  77  Tr: Hi 4 ] n  means dan  signific and kg J B  'to carry on the shoulder.' It consists of rin  A  'person'  phonetic.  Analysis L i Xiaoding tries to separate them together in the composite graph  ^  from ^  , which is curious because he lumps  , both of which he identifies as He > ] tt  'the Yellow River' (JWJ 11.3261). He finds only one complete example of the graph (Qianbian  ^  7.1.4, though personally I find this rubbing too dark to read), and one  incomplete example, ^  (Cuibian 543), which he reconstructs as  neither. L i identifies ^  as yin i&  A Shima lists  because of its similarity to the seal form ^  5b. 10b), and says it is the primary form of dan  (SW  'to carry on the shoulder.' He  explains that it depicts a person carrying something across both shoulders, whereas  ^  shows a person carrying something across one shoulder. If he were right, one would expect the meaning 'carry' to crop up in one of the phonetic compounds of yin ^ does in the case of hi  , as it  , but we do not find this (see GSR 656). This is why he is  driven to propose that it is the primary form of a word from another phonetic series, dan . This is fairly reasonable phonetically since the yin rhyme and the zhan )%  series is in O C i£n  phonetic series is in O C gjn rhyme, and these two rhymes  differ only in that the former has a close vowel while the latter has an open vowel (according to Pulleyblank's reconstruction, and in fact also Karlgren's). However, the yin series gives strong evidence of being what Pulleyblank (1984:170) calls an *l-type initial, while the zhan  series would appear to be a *t-type, so these two series are not  really that close. I think that the graphic similarity between 'i  and yin  is purely  fortuitous, and that the original form of yin 'fcj probably depicted a person lying down with his head on a pillow, i.e. the primary form of zhen  'pillow.' Although an earlier  !ln both cases, he says, it is a person's name. In his commentary on Cuibian identifies the graph as h£  (/f »J ), says it is the diviner's name.  543, Guo Moruo, who  78  form of this character has not been found independently, it is perhaps preserved in the bronze forms of yin 'fZ> and shen >/£J :  (O/NJWB 62/72)  (O/NJWB 1427/1824)  If the bone form of y i n iZ> existed, I would classify it in Part 3 of this chapter, where ^  represents a person lying down. The present graph shows a person carrying an axe over the shoulder. As L i rightly  points out, the axe is the primary form of kg fa] k'aa < *k'al 'axe-handle' and acts as a phonetic hint in h i ^"T -yaa' < * g a l ' . He also points out that in the rare variant (Xucun 1.637, where it is actually the name of a. fang ge_ ~\  kwaa < * k a l 'spear' phonetic.  ^  tribe), the person is carrying a  This does not make quite such an accurate  w  phonetic as the axe, due to the labiovelar rather than plain velar initial. In later times, M  ' "J  was reserved as a loan for the WH-word h i , and h i  write the word h i 'carry.'  4**3 'lotus' was borrowed to  The process of borrowing and re-borrowing is rather  complicated. H i ^ *) derives from ^ \  through the separation of the 'person' and 'axe-  handle' elements and the addition of a kou  a  'mouth' element, which may have been  added as a sign of desemanticization for the WH-word usage but now forms an integral part of kg "J •  Usage It seems to be the name of a place or perhaps of the people living there, There are no examples of it in its primary meaning 'to carry.'  79  72 ^  JGWB 359: \ \  , &  = ji &  , 3598 (unidentified)  S.14.1x106  JWJ 3.915: ^  , k  SW3b.9a: £  £  &  =J1  ^  jHl . ^  SWDZ 3b. 18b: Ji &  . &  : *  2-  ,  ^  .  1  : "t  :* i  * £ &  i  1  ;^']  &•  and d l i £ ^ gloss each other in SW. The second old form occurs  in the character fjn Rj . The last old form consists of chuo >1_ and probably bj Tr: J i ^  means djti. ^  'to reach.' It consists of y6u  .  'hand' and ren  'person.' [Xu Kai adds the interpretative note here: &  "j^O A - ^L> 'it represents  reaching the person in front.']  -this is how it is written on  Qin steles.  ^  is an old form of j i  is another old form of ji  *  sy  e  t  mother old form of ji TyL .  The first two old forms do not look like complete characters. The last old form is obviously a variant of djli  , and I am surprised that Duan did not realise this.  Analysis The graph shows a hand reaching a person from the rear, thus symbolizing the meaning 'reach.' By extension it also came to mean 'as far as, up until, by the time' etc. Both these usages are evidenced in the bone inscriptions. JWJ and JGWB both include the graph ^  as a variant, but S separates it, listing  it after the present graph. Since it occurs only as a person/place name, this identification cannot be verified. Note further that there are no unambiguous examples of person/place name.  Usage  ^  as a  80  JI ^  is used in two main ways:  1. In expressions like ^  ^  <® \ \  'when it comes to the present fourth  month it will rain,' where it clearly functions as a time adverb. 1  2. In military inscriptions, where it may be interpreted as 'catch up with, reach (the enemy),' e.g.  (Zongtu 22.4) Jiashen-day cracking, tested: Cha will reach the Xuan tribe.  100  JGWB 407: 4>L  ^  (not in SW)  25.4x9  JWJ 3.1027: 4sL  SW3b.l3a:  (notinSW)  &  &  M  §L  ,  M  4  J X  : it  £  <J  /A  A. Tr: Y i <f  means 'to guard the border.' It consists of shu  walking radical], q >^ : the old form of y i 4 h a s rdn  'lance' and chi  \  [a  'person' [instead of chi  Ai X u Xuan says that chi  ^  trMajk is also phonetic in y i ^ _ j w i a j k .  However,  although they happen to be in the same E M C rhyme, they do not come from the same O C ^Translation of this phrase suggested to me by Professor Takashima.  81  rhyme (du6  ffip  *-ak and xi  fy%  reconstruct in the case of yi  *-ac respectively), and their initials, though hard to  , are unlikely to have had much in common.  Analysis The present identification was made by Y u Yongliang (JWJ 3.1027). Although SW gives /| 5^ authenticity.  as the old form of y i 4!L  He says that y i ^\sL  signifying 'march,' and shu.  , L i Xiaoding refuses to recognize its  is a semantic compound consisting of chi ^  , a weapon, thus its primary meaning is xrngyi ^  'to go on a military expedition.' He says that the bone graph  , ^  , on the other hand, An  shows a person being beaten, so it cannot have any connection with yi IX- , and its meaning, he claims, must be something like puji  'hit.' However, this is not Art  necessarily so. Another common meaning of yi 'ISL is as in shiyi (someone to do something),' and the Yupian defines s M  i± Art  as y i ^ s t  '\9^  'to cause  (though such a  late work should not of course be relied on as one's primary evidence). This could easily be its primary meaning. The bone graph could be interpreted as showing a person being forced to do something, 'beaten into service' one might say. 'Military campaign' could be just a specialization of this meaning, but it could be this specialization that led to the /  replacement of the r£n S**. element by the walking radical chi that binds the various uses of yi  Aft  A  I . The semantic thread  together is 'obligatory service.'  Usage The inscriptions are rather hard to interpret, but I think that in most of them it may stand for y i  'plague,' as Rao Zongyi (1959:520) suggests. This usage helps to An  confirm the correctness of the identification as yi 1  .  82  108 ^  JGWB 740: xiu. \%~ J  S.26.3xl8  JWJ 6.2023: xiu  SW 6a.23a: Tr: Xiii  means 'to rest.' It consists of r£a  'person' leaning against mil  : xiu. ^V%- is sometimes written with van f~  'tree.'  'roof.'  Analysis The bone graph also consists of 'person' and 'tree,' so the identification seems not unreasonable, though since it is always used as a place name one cannot get any contextual support for this identification. The character xiu.  that SW gives as a variant also has  the extended meanings 'shade, shelter, protection.' If the identification of the present graph is correct, then there seems to be no reason not to accept the SW analysis of the seal graph, that it represents a person leaning (resting) against a tree.  Usage Place name.  109  * f  JGWB 3993 (unidentified)  S.26.4xl  JWJ 6.2023: xju  $~  83  SW6a.21b: &  , 3#  JflL  . kk  &  , Ik /K  means 'to pluck.' It consists of mil  .  'tree' and zhao  'hand.'  Analysis L i Xiaoding includes the present graph, with reservations, as a variant of the preceding graph, since it is also used as a place name. However, I think it is more likely to be a variant of cai % I ^  (JGWB 737). It shows a person plucking at a tree, and at the  same time the % element could also be a phonetic hint. Notice that it is not the same as mu ^  /  . It occurs commonly in the graph  , which stands for a time word in  the inscriptions. Yang Shuda (JWJ 6.1967-9) observes that the ^ \  , a variant of a i *f  , \  'to attack' (JWJ 12.3777), where it would appear to have  the same phonetic function as cji later written zai ij^  element occurs in  I ~k . H e suggests that  'year.' It seems possible that %  represents the word  is the original pictograph for zai  'to plant' (GSR 943y), which the Ming dynasty work Zhengzitong JE- ?  lit  defines as 'sapling.' The focus in this graph is actually on the hand, and indeed it is only the hand that is left in the usual form % . The graph as a whole may be transcribed *  Usage Place name.  Ill  JGWB 3643 (unidentified)  S.26.4xl  JWJ 3.1083:  (not in SW)  •  84  Analysis The graph shows a person, whose hand is drawn in, with a stick, and there is another hand at the bottom of the stick on the other side. It is not clear which hand is holding the stick and which is resisting the stick, or whether both hands are holding the stick.  1  At any rate it seems clear that somebody is trying to beat somebody else with a  stick, and some such meaning as 'attack' fits very well into the OBI context:  (Shiduo 1.415, ap.S.26.4) Yihai-day tested: Geng (?) should not attack the Fang (tribe).  Here again the focus is on the hand, and indeed one of the two contestants in this struggle has only his hand drawn in. However, the rin  \  element still suggests that it is  a human activity, and the OBI context corroborates this.  114  Jy  JGWB 571 :jf  H  S.27.1x23  JWJ 4.1549: ji  SW4b.21a:  i L It.  IS , £ «  ^ . t % fc-J* M , tkH  W * , t f .  iprofessor Takashima has pointed out to me that the graphic intent of the present graph, as I interpret it, is much like zheng W  / -Jp- 'struggle,' i.e. two hands holding clubs hitting each other.  85  S W D Z 4b.53a: JI 4^3  occurs with the meaning 'to plough in various classics, with 1  reference to the emperor's ploughing ritual. The commentator Chang Chenzan explains that the emperor had to ji£ I e  'borrow' the people's labour because he could not finish  all the ploughing by himself. Tr: JJ 4^  refers to the emperor's ploughing one thousand mou. In ancient times, the  people were employed as if borrowed, and so it was called ji  [which is phonetically  similar to the word for 'borrow,' as one can see from the fact that they both contain xi 1? phonetic]. It consists of lei  'plough' signific and xi  & phonetic.  Analysis The bone graph shows a person pushing a plough.  The r6n \  element is  augmented by two hands and a foot, so it really conveys the idea of a person struggling to push a plough along with his foot pushing it into the ground, and its usage in OBI confirms that it means 'to plough.' This meaning survives in only a few places in soft texts, where the character is written 4^  (i.e. augmented by the grass radical), and it seems to be  nominal, referring to the piece of land that was ploughed, rather than to the act of ploughing itself, e.g.  (Liji.jiyi,Wt  24/31)  "Thus it was that anciently the son of Heaven had his field of a thousand acres." (Legge 1885.IV:222, % 5)  The commonest usage of the character  in soft texts is as a loan for the word j i £ 'to  borrow,' which is now usually written  . X u Shen is at pains to explain this usage,  but all he ends up by doing is making a pun.  86  In bronzes, whether as an independent character or as part of another, ji always has 2d ig  added as phonetic, e.g.  3  (O/NJWB 569/701)  In the modern character, the plough has been stylized and the ploughman has vanished, so all we have left is a phonetic compound, which is how SW perforce analyses it.  Usage Always used in its primary meaning 'to plough,' e.g.  fcZY  . J : £ * # ** . M i . + - A • ^  (Houbian 2.28.6)  Gengzi-day cracking, tested: The king will perhaps inspect the ploughing. (He) should go. Twelfth month.  Guan jj^ here could perhaps alternatively be understood as guan >|f guan . | | ~ 'to pour a libation on,' and ji  'to irrigate,' or as  could be understood as a noun 'ceremonial  field.' SW (lb. 19b) reserves the nominal meaning for the character j | | L B , defining it as jiji  7  ^ %Q 'the ceremonial field.' Karlgren (GSR 798b) defines it as 'field ploughed by  king whose produce was used for sacrifice.'  293  JGWB 1.116: w£i %^ 'person's name'  87  S.46.1x3  JWJ 9.2909: w i i  SW9a.l6a: J l  Tr: Wgi ^fL hji  , &  . M  'fearful' means i  &  , ht  |  A  fI  ^  'horrible.' It consists of fl 1  £  &  ^  3"  | ,  'devil's head' and  'tiger' abbreviated. Devils' heads and tigers' claws are fearful.  ^  : the old  form is abbreviated.  Analysis The graph appears to show a devil wielding a stick, which is indeed fearsome, but the gui *f  /  element is of course phonetic. This clever way of making the phonetic  serve also in a signific capacity is typical of the way in which the creators of this script designed their graphs. Rather than the usual form of gui  , which is  ^  (see  III.l.i.283), what we actually seem to have here is a member of the Gui tribe (see I.l.i.a.292)  The original meaning was perhaps 'to threaten' and it is undoubtedly  etymologically related to the word wei ^ In soft texts, wei -^L  'awesomeness.'  usually means 'to fear,' but the graph appears rather to  show the act of threatening. This prompts me to investigate whether this character could ever have had the meaning 'threaten.' Note that the word w_ei f$k\  uj, from which wei  ' u j is undoubtedly derived, does not mean 'fear' but 'fearsomeness, awesomeness,' h  which can perhaps be understood as 'threateningness,' and in bronzes we find the usage of these two characters somewhat confused. For example, in the to  find the expression 1  inscription we  -r ffl "fL.  'fear Heaven's awesomeness,' in which the character  Or perhaps it should be read w i 'to loathe.' Karlgren gives this meaning for w£i J^?^ in his Gloss 1000.  88  wei -jj^ is used for both words (in the Shujing, 'Heaven's awesomeness' is written J$L ). There is also a common bronze expression yyjayj A authority,' which on the bell /  &  is written with  {=  ) 'demeanour of  (i.e. wei  augmented  by pjx "^C ), and a similar phrase meaning 'authoritative and circumspect' is written *  &  onthe  4> %  ^  but  onthe  In the Shujing there are a couple of examples of wei  . being used as a verb  where it may be understood as 'to threaten, to cause to fear':  ^  |'l  i  • (Tongjian 5.293-6)  "If not, one overawes them." (Karlgren 1950:11 fll4))  f%&$L,%^%7k&  • (Tons/Van 16.818-26)  "Do I overawe you? (No), by this I take care of and (nourish) sustain you all." (Karlgren 1950:24 ^26)  In these two cases, wei  has a direct object (in the second case cliticized due to the  interrogative nature of fli _g_ ), so it is definitely a transitive verb. Karlgren translates as 'overawe,' but I think it is simpler, and more appropriate to the context, to understand it as 'threaten.' In the second example, Pan Geng is trying to cajole the Shang people into moving to Yin, and the tone of his whole speech is very threatening. The tone of the above quoted passage would perhaps be better captured in a less literal translation such as "Surely you don't think I'm threatening you? I'm doing this for your own good!" Pan Geng is trying to say "I am not threatening you, I am persuading you" (though he is of course threatening them). In spite of the Dai Kan-Wa Jiten's claim (7.1088 and 3.700) that both w M ^ wei /lie  and  can have the meaning odosu 'to threaten,' I have not been able to find an  unambiguous example of wjsi -t?<_ in this usage. The most probable example given there  89  is perhaps the one from Liezi.Huangdi: ^  ^  . This could be understood as  he neither threatens nor gets angry, but the interpretation 'fears' is not precluded. I think 1  the context supports the meaning 'threaten,' as the passage is talking about how the sage does not manipulate or coerce people and yet they still do his bidding, and 'threatening' seems to be more appropriate to coercion than fear does. However, the ambiguity is still there. Graham, for example, translates "He inspires no awe, he is never angry" (1960:35), whereas Wilhelm translates "Sie wissen nichts von Scheu und Zorn" (1911:2). Apart from the uncertainty of the interpretation, the date and authenticity of the Liezi are also a problem, so it is not a good source for determining ancient usages. My conclusion is that the graph  was originally devised to portray the concept  denoted by the word a^i 'to threaten' (verb) or 'threateningness' (noun), but was also used for the qusheng derivative wj^i 'to fear.' Later, the character -p<_ came to be used only for was borrowed for the word wei 'to threaten, threateningness.'  1  Usage Unfortunately it occurs only as a person's name.  J G W B 1071: xian 7CJ  S.75.1x234  J W J 8.2809: xian 7 b  According to SW (12b.3a) the primary meaning of wei  is gu.  'husband's mother.' It consists  of n i 'woman' signific and xu r% phonetic. Duan Yucai (12b.7b) suggests, in all seriousness, that the meaning 'awesomeness' is derived from the meaning 'mother-in-law.'  90  SW8b.4b:  , ^  j£  jfej .  kk  )L  , kk 2-  .Y\  ...  Tr: Xian yC, means 'to advance.' It consists of )l> [= positional variant of rin 'person'] and zhi  K  'to go.'  Analysis Somewhat unusually, the SW analysis happens to be correct, though not terribly insightful. X u Xuan tries to explain the relationship between the two elements by saying 'to go above people, this is to advance,' but this smacks of sophistry. Furthermore the bone graph sometimes simply has zhi tr  / i t  'foot' on top.  I think the basic idea may be something like 'to precede, to go on ahead,' and this is conveyed by 'foot' on top of 'person.' The foot is welded on top of the person as if it were his head, and this is probably a conscious piece of graph design. JGWB and JWJ also include  ^  as a variant. This graph has sheng %. as the  top element. L i Xiaoding claims that sheng  I  and zhi )£  I  are often  confused in OBI, but it seems odd that the same graph could be written with either sheng %.  or zhi ^— • I believe Shima is correct in separating % , along with 7  is used mostiy as a person/place name. ^ 'advance,' but  , since it  occurs quite clearly in the meaning 'first' or  never occurs in this meaning.  ^  1.18, where Zhang Bingquan transcribes it as l|o. ^ same plastron (inscription 1.17), we find xian  occurs in the inscription Bingbian in his kaoshi. Significantly, on the . It occurs as the object of hub  ) 'to capture,' and would appear to be the name of an alien tribe. The same plastron talks about attacking the Bu ^  , presumably also an alien tribe, and  Professor Takashima has drawn my attention to the following Zuozhuan passage (HY 340/Zhao 1/3 Zuo), which Zhang Bingquan quotes (kaoshi p. 15):  91  Shang [had] its Seen and P'ei [i.e. tribes that caused them trouble].  1  Bingbian 1.18, on the other hand, I would translate thus:  Bingyin-day cracking, Zheng: Call on old Marquis Zhuan of Long to kill Pei.  U s a g e In the meaning 'first' or 'advance,' but also the name of an alien tribe in Bingbian 1.17.  655 ^  .  f  S.107.1x58 661 "J  JGWB 1072 (all forms): jijn  fj  S.107.3x76  JWJ 8.2811:  T , X =ji£n  JWJ 8.2685: f  SW8b.5a: |  (Jinghua 10.9) = gin  , ^  ^  . /A  >»L , Al  Tr: Jian ^> 'to see' means shi rin  'person'] and mn S  [S. 107.2 lists as Jinghua 9.9]  g  . &  ...  'to look.' It consists of 'eye.'  iTranslation fromLegge 1861.5.2:577.  [= positional variant of  92  SW 8a. 15b:  S W D Z 8a.42b:  9  other.' Hen  . means ^  Tr: Gen  t  means 14"  @ ^{i  'angry eyes confronting each  'disobedient.'  'perverse' means hen ^ f c 'disobedient.' [This is a punning definition.]  The graph consists of bi t  'to line up' and nn± @  'eye.'  'Lining up the eyes' is  intended to portray the idea that the eyes are on the same level, the one neither higher nor lower than the other [i.e. in mutual opposition]. The still."  Notice how gen  1  consists of hua  ^  f<L consists of M t  says: "Keeping his hips  Yijing  and mu 9  , whereas zhen lj|  and m i @ .  Analysis The graph consists of 'eye' on top of either 'standing person' or 'kneeling person,' thus indicating the meaning 'see.' The disproportionately large size of the eye shows that this is the focus of the graph. Shima separates  from ^  , but JGWB and JWJ  combine them as jian 7U . It is hard to find contexts that prove that  same character, but since both forms occur as jian  |L  T  ^d %  are the  in bronzes (see O/NJWB  1170/1442), it seems likely that they are indeed both to be identified as jian  .  Professor Takashima has also drawn my attention to the interestingly parallel phrases  T Jlj  Waibian  5  6p  (from  Yingguo  1784) "see the yjn of Zheng' and ^  34) 'see the yjn of Fang' (the exact nature of yjn  tP  £p  (from  is unknown, but it seems to  refer to some kind of human). The form with the staring eye ^  is separated by Tang Lan (and followed by L i  Xiaoding) on the basis that the eye is looking back. Tang also suggests that yjn  iHY 32/52/3. Translation from Wilhelm 1967:202.  'eye'  93  is an accretory form of gfoi j j | , while admitting that the meaning is different. Since the 1  inscription that  V  occurs in is fragmentary, it is hard to say whether it really is not jian  IL • JGWB 1072 lists another 'backwards looking' form ^ (Jingjin 2104 ), but the 2  inscription is hard to make sense of. ^  4  ( kit  lc  ^  3  L i Xiaoding (JWJ 8.2686) also cites a bronze form  )> but it is part of a person's name, so the context provides no  support for this identification. O/NJWB does not have gen  f^L as an independent graph,  but it may be found as an element in the bronze graph for xian ap. O/NJWB 1814/2321). Tang explains that the original meaning of gen  was fangu  'to look behind.' However, this meaning is not attested. It occurs to me that the emphasis in the graph ^ like bin  is on the enlarged glaring pupil, which could convey a meaning  'hate' or hgn  'fierce.' Cf. Duan Yucai's comment quoted above about  'angry eyes confronting each other.' Although he could not see the bone graph, his intuition may be right. On the other hand, perhaps it is simply the primary form of yan gg, 'eye,' as suggested to me by Professor Pulleyblank. The present graph is a very rare example of the standing and kneeling elements being interchangeable in the same grapheme. 'Looking' is an action that can be performed in a standing or a kneeling posture, but I think that the use of the kneeling figure in  ^  was perhaps inspired by the usage 'see, visit (a superior),' which is attested in OBI:  a  v- , i & i : 4 * * . I  3-  .- H  ^However, as Professor Pulleyblank has suggested to me, if the words were close enough that gM could be phonetic in yM  , it would still make sense to say that gen  for van §8. and was a phonetic loan for the word JGWB misquotes the number as 2014.  &  .  2  T"fc(7>£:£fcjE.(7)*L. S does not list this inscription under or under yi f | . Given in O/NJWB Fulu 2.30a.6/2.385 as an unidentified graph.  4  was originally the graph  94  (Bingbian 124.15) Jiwei-day cracking, Que tested: Fou will perhaps come to see the king. First month.  Other inscriptions suggest that Fou  was probably the head of the Ji ^  /  tribe  (see Bingbian 302.1/2). I suppose that the status of the visitor did not have to be literally inferior to the visitee.  The kneeling element may simply suggest the respect that  traditionally the visitor has always shown to the visitee in Chinese culture. Shang Chengzuo (JWJ 8.2811) suggests that we compare jian ^ '  /* 3i. ( S~ ), and this is indeed very instructive: jian ^  looking forward, whereas wang  1  with wang  simply shows the eye  , according to the explanation proposed by  Takashima (1989:114-117) shows the eye looking backwards, thus suggesting that it is looking for something that it cannot see (see discussion under next graph). Significantly, there is no * ^  : the kneeling position is not appropriate for conveying the idea of  looking about.  Usage  Mainly in its primary meaning 'to see,' and extensions of this, but in some cases it may also be a person's name (especially the 'standing' form).  690 $  .  <f  JGWB 1035 (both forms): wang  S.l 10.2x195  JWJ 8.2711 (both forms): wang  95  SW8a.l7b: ! ?  £  .  i  ,  fl  ;i  #  ,  *  f B }|  SWDZ 8a.46b: [Has wing ^  0 *8  JL.  £  .  vk  $R  §.  I :4* $ * .  . Ik ft ,  , and emends yi V"A to 51 ^k  instead of wang fjf  Tr: Wang 3_ : when the moon is full, it wang  M  .]  'faces'the sun. It is at this time [i.e.  the middle of the month] that one pays court to the sovereign [or according to Duan's emendation 'the moon facing the sun, and thereby being fully illumined, is like a minister paying court to his sovereign']. It consists of yue r\ ting  ; ting  wang  SW1  'moon.' chen  'minister,'and  'court.' jg.  : the old form of  here stands for chaotfng  is abbreviated [it has no 'moon' element].  2b.l9a:  . £ "t & ^  S W D Z 12b.46a: Wang  £  £ it  & . «. " t , £  %f .  and wang J L . are clearly different characters, though now  often confused. Tr: Wang _i» means 'to go outside and look for someone's return.' It consists of wang T— 'eo awav' signific sienific and wang wane S~ go away' . i - abbreviated phonetic.  Analysis X u Shen tries to separate wang  and wang 3 1 , putting the former under his  ting - - radical and the latter under his wang *— radical. However, in spite of Duan's x  protestation to the contrary, it is clear that they are variants of the same character, both denoting the same basic word. In bronzes we find J L ,  (O/NJWB 1118/1378) and  (O/NJWB 1621/2059) all used in dates to denote the phases of the moon. The bone graph consists of ch6n  on top of r6n  f  . The character wang  thus  cannot portray 'gazing up at the moon,' as L i Xiaoding maintains, though this romantic 1  1  BL£  L i also explains that the reason why wang ^7 bright to look at! 1  contains 'moon' rather than 'sun' is that the sun is too  96  notion is rather appealing, and has been exploited by poets, as in L i Po's famous line f !i JE-. ^  ^  H  • Although it is possible to regard the moon element as signific, H  indicating the usage as in wangyue  'full moon, an alternative possibility, 1  proposed by Takashima (1989:113 ), is that the graph identified as yue 1  H  was  polyphonic, and acts here as phonetic, as is supported by such other characters as ming and meng jm , but when this other reading of yue " wang  had been forgotten, then  was added as phonetic. The replacement of chert !aL by wang ~C  in the popular form ^  was perhaps  inspired by their graphic similarity, and is probably primarily of phonetic intent, but notice also, as Takashima (1989:114) points out, that the meaning of wang " d exist,' is also semantically relevant, as wang  'disappear, not  denotes "the act of searching for  something that has disappeared or does not exist at present."  Usage It occurs largely as a person's name, especially in the name of the military 3^- . Wang is perhaps the place name or tribe name, and  commander Wang Cheng  Cheng his personal name. We also find the name Wang Yang (?)  jf.  (S. 111. 1x4).  Wang occurs as a place name in five inscriptions (S. 111.2: Xubian 3.29.2, 3.31.4, 3.29A, Houbian 1.10.3, Jimbun 2878). In a few cases it could be the name of the sacrifice offered to mountains and streams which is mentioned in the classics, and which was so called because the sacrificer looked at them from a distance, e.g.  i  :  >V  &  3  L  &  £  •  (Jingjin 1347) Tested: It should not be the king himself who performs the wang sacrifice.  *I am grateful to Professor Takashima for sending me a copy of this material before its publication.  97  (Buci 770) Tested: Perform the wang sacrifice to Mount Hua.  715  JGWB 1403: 7u  f  1  (notinSW)  S.114.3xl  JWJ 12.3550: %J (notinSW)  SW 12a.8a: S W D Z 121.17a: [Explains the function of the d £  element with the pun 'the ear (££  'gets' something.'] Tr: Ting ^ % means ling  'to hear.' It consists of gr  fwhich I attempt to explain below!. with ting i~  'ear' and d6  signifies  phonetic.  Analysis The graph consists of 'ear' on top of 'person.' L i Xiaoding does not identify it with a modern character, but comparison with jian primary form of ting 3%  'to see' suggests that it is the  'to hear.' The present graph may be identified with the ~%  portion of the modern character. The rin \  4 follow Guo Moruo'sfirstidentification of ^ 9). Guo identifies the ^  /  element was phonetized to ting i  , rather than his second as y j £ -jfj  element as the SW (6a. 15b) character hui $  /  . The  (see JWJ 9.2918-  .defined as  'two-bladed ploughshare.' Duan Yucai (SWDZ 6a.41b) points out that in the Fangyan (5/35/27) this word is written  . It seems possible that, rather than depicting a two-bladed ploughshare, ^  stylized flower, i.e the bone form of Jmi  , for which the SW (6b.3a) seal form is  is a  98  addition of the d£ ^  element is mysterious, especially since the original meaning of d£  itself is a moot point. The 'heart' element suggests something to do with mental perception, cf. cong  'perceptive, intelligent' also contains it (the 'ear' element in this  character is probably a later accretion). Duan Yucai is probably on the right track, and the de" jg- element perhaps indicates 'perception,' which may be seen as the abstract correlate t o d l ^f-  'to get (physically).' D £ %• refers to a mental 'getting.' $Y ^  defines it paronomastically as  ^> A  [ ^ ) ^-f*  SW (lOb.lOb)  tL  'outside  getting from people and inside getting from self.' This definition is admittedly not very intelligible.  Duan Yucai (10b.25a) says that  enrichment ( -fc  -If  M"  J  ^%  ) while  ^•f)  MlP  ( r & " ~ % )•  h  1  1 0 3 1  ^£ '•Vsr and d £  /S  refers to spiritual ^f* ^  ^  A_  refers to largesse  probably originally referred to the mental  and physical aspects of the same basic notion of 'getting.' D £  / ' i f - , with the hand  clutching at pecuniary wealth, clearly indicates a very physical, tangible kind of 'getting,' while the 'heart' signific in d £  indicates mental getting, i.e. perception. The verb  'get' in English also has both these meanings, e.g. 'get wealth' versus 'get a joke.' It seems probable that d £ 0 <i- and d £ 0  have always been homophonous, and Chinese  scholars have long felt them to be conceptually related. They are most likely the same word. The basic concept may be defined as CEPTION. In a language like Latin, a wide variety of prefixes are used to distinguish different aspects of the same basic concept-thus we have the concrete notion of CAPTuring as opposed to conCEPTlON, perCEPTION etc. on the abstract level.  In ancient Chinese, by contrast, one has only the unadorned  monosyllable to express this polysemy. Thus I surmise that the word d£ originally referred to any kind of 'ception,' but that the concrete and abstract aspects were distinguished in writing. In bronzes, the graph for ting O/NJWB ( ^  1507/1924).  '\ K ^. !  V  has, not d£.  However, the examples  come  , but gji &  I v£  from a single  ), so one cannot know if this was a standard form.  (see  vessel  99  JGWB includes various other graphs under his ^ instead of a standing figure j |  , one of which has a kneeling  , but the lack of context makes it impossible to say  whether they denote the same word or not. However, the situation could be similar to the one we have seen with jian  . Note that ting  has a pingsheng reading with the  meaning 'listen' and a qusheng reading with the meaning 'hear' in the sense of 'obey.'  1  The difference between listen' and 'hear' is parallel to that between look' and 'see,' the former of each pair implying active volition and the latter involuntary reception. One could suggest that the standing form ^ kneeling form ^  simply shows the act of hearing or listening, while the  was inspired by the sense of 'obey.'  Usage Unfortunately the graph has no context My identification as ting  is based on  / C> and me structural identification with the %  the parallel structure of jian  part of  ting J ! - .  801  Not in JGWB  S.136.2x2  Not in JWJ  1  GSR 835d defines the pingsheng reading as 'hear' and the qusheng reading as 'listen to, acknowledge,  obey.' The Guangyun gives the meaning ling gives the additional meanings dM 4^ edn.) only gives the definition ling shenwen  | ^  'hear to both readings, but to the qusheng reading  'wait' and m6u  1  'consult.' Wang Renxu (Long Yuchun 1968  to the pingsheng reading, and defines the qusheng reading as  'consult.' According to Long Yuchun (1968:555), the wen Itt^  written with the character the meaning.  in this definition is  in another copy of Wang Renxu's Qieyun, but this does not seem to affect  100  Analysis The graph consists of 'person' signific and jri£. ^  /  phonetic. From the  context, all one can deduce is that it is a verb (it occurs after jvit ^0  and al ^  )•  However, since 'human activity is the chief significance of the 'person' element, one 1  might suggest that it represents the word jig  1847  p*j]  'connect, come in contact' (GSR 635e).  JGWB 4174 (unidentified)  S.273.3x62  JWJ 3.1063:  ^  SW3b.l7b:^ , J|  JSU .  U  Tr: Kou 7&> 'to rob' means bao ^  , kL  %  'to do violence.' It consists of p _u_ gu_  $L 'to hit' and  wari 7 0 'completely.'  Analysis In his commentary on SW, X u Kai tries to explain how pu. ^ _ 'hand holding stick' and wan  'complete' act together to represent the meaning 'rob.' However, as  one can see from the bone graph, the original representation was of a person wielding a stick inside a house. Karlgren (GSR 111) lists the early meanings of kou jQj  as 'to rob,  robber, invader, bandit,' and I think that 'rob, robber' must be the primary meaning. Why else would a person be wielding a stick inside a house unless he were robbing it? Ye Yusen (JWJ 3.1065) says that the dots indicate the mess made by the robber (as he rummages for valuables, presumably). He also describes the evolution of the graph by  1Q1  reference to bronze forms, in which we find the 'hand holding stick' behind the 'person,' thus illustrating a person being attacked in his own home:  (O/NJWB 427/533)  In the earlier forms, such as the first example I list above, the 'person' is written j this evolves into yuan 7 0  , but  in the later forms, as in the second example above. This is  probably a phonetization (for the phonetic implications, see Pulleyblank 1963:208). NJWB 1230 lists a graph  \$L_h  ) which it transcribes as '[% (not in SW), but  this should probably also be identified as kou x£j .  Usage It occurs in two main contexts:  1. In the expression Duo Kou  % 'fjij , probably a military force comprised of convicts.  2. As sacrificial victims.  Other inscriptions refer to zhi 4fy (It  ) 'killing,' and  ft  'shackling,' \^  I* %$Hj 'beheading (?),' sha  ^  'amputating the feet of (?)' the k6u H> . These are clearly  punishments, and suggests that k6u 73, refers to robbers who have been caught, i.e. convicts. The soft text term sikdu ^} 7&  ^  ^  a  v  em  *  s m  o  r  e  for 'Minister of Crime' suggests that kou  general meaning 'convict, criminal,' rather than just the specific  meaning 'robber,' and the bronze graphs could perhaps be re-interpreted as showing  *Not in OJWB. It would be between 997 and 998).  102  punishment being inflicted on a convict in a prison. In classical Chinese it is also used as a derogatory way of referring to enemy peoples.  2120 \i  JGWB 240:  (notinSW)  S.323.1x27  JWJ 2.609: xing  ^  SW2b.lla:^  , >C  S W D Z 2b. 18a: Bu ^  £  jfi  means xing  j & . >M fa  , JX T  'to walk' and qu  • J%  -  means zou ^  'to run.  Together they refer to going by foot in general. Tr: Xing ^  refers to a person's walking and running. It consists of chi  paces' and chu  2  X  ^  'small  as variants of xing ^  /^J".  1  'stopping in ones's tracks.'  Analysis L i Xiaoding includes the graphs \ "\\ and ^ "\  They differ in that a 'person' element is added, and in the present graph the xing  ^  element is also abbreviated. His classification seems reasonable on graphic grounds. He makes the interesting distinction that ^r- is the noun hdng 'road,' while  , which  shows a person walking along the road, is the verb xing 'to go.' Xing differs from hang in having an *-r- infix: harig -yaan.  < *gan,  xing  -jra j n. < * g r a n . .  ^his pronunciation is based on the Tangyun reading Guangyun reading is the same (see Shen 1960:361). 2  r  Jji  "tJ)  This pronunciation is based on Xu Shen's sound gloss 'read like c M ^  However, it is hard to tell  given at SW 2b.8b.  The  ' (SW 2b. 10b). The Tangyun  reading given there is "Q* - i - * "t/0 , and the Guangyun reading is the same (see Shen 1960:361).  103  from the inscriptions how <ifr is being used, while there are examples where  can  quite reasonably be interpreted as a verb, e.g.  E £ :£  6 £  • (Yibian 947)  Jichou-day: The king will not go from Qiao.  T e.  <£  ) :  ,jx  is, ? M - f . (Yechu 1.30.4)  Dingsi-day (tested): (There will be) small rain, (we will) not go.  I suspect that the Shang used the same graph for both 'road' and 'go,' as in fact has been done ever since. The identification of  as xing ^"T is supported by the occurrence of  this graph on the Stone Drums, as Luo Zhenyu and Qu Yipeng point out, where the rhyme scheme suggests that it is a variant of xing  It seems best to regard  as an old  variant of xing ' T j that died out, but one may still maintain that it was the concept of human action that inspired the addition of the 'person' element in the first place. As for the present graph, there is an example of it occurring in the same inscription as \ \  (albeit as  a person's name) which strongly suggests that it is a different grapheme:  (Ci«Wan51lA)  Xinwei-day cracking, Hang tested: Perhaps summon ^  to go, have (occasion  to) meet with.  t h e r e are two rhyming occurrences, and in both cases it rhymes with O C yang xlfy rhyme words. See Mattos 1973:IV.5d-g (discussion p.329) and V.7b-e.  104  As usual, there are a number of different ways in which this inscription could be interpreted. I have given only the most obvious reading. At any rate, it would seem preferable to identify the present graph as either yong ^C. o r p l i Jjk-  (JGWB 1352 and  JWJ 11.3411 resign themsleves to not being able to distinguish the two, and the problem is compounded by the fact that they are used mainly as person or place names). Whether one identifies the present graph as xing ^"T or yong ^ K - 'to wade' (of which Karlgren suggests yong  'to go,' pjri. ^k- 'to send,' is the primary form-GSR  764k), all three refer to human actions. The graph seems to be huiyi ^  ^  in nature  (to use X u Shen's term), which makes it unlikely that it was created specifically for the person's name that it often stands for in the inscriptions.  Usage Mainly as a person or place name.  2234  JGWB 1007: f i 4"%  S.329.4x713  JWJ 8.2657: f i  AX  SW8a.l2b: ffi , ^  . Kk  k.  &  . —  &  %  k> .  SWDZ 8a.34b: [Quotes various classical commentaries to show that ji ^ the primary meaning, and that zhengfa' ^iE. 4\  'to strike' is  'to attack (in war)' is an extended  meaning.] The Gongyang commentary says: "In the Chunqiu, the word fa 4\  is used  both of the attacker and the attacked-how can this be? When it refers to the attacker, it is read long. When it refers to the attacked, it is read short. This is Qi dialect." The present  105  rusheng reading is the short reading...while the qusheng reading [in certain places in the Zhouli ] is the long reading. Tr: F i /[% means ji ^  1  'to strike.' It consists of 'person' holding 'spear.' One source  defines it as hM I^C 'to defeat.'  Analysis SW analyses the seal graph as a person holding a spear, but the bone graph shows that originally it depicted a person with a spear running through the neck. It thus depicts the act of striking an enemy dead. The second definition given in SW is ambiguous, since with a voiceless initial means 'to defeat' but with a voiced initial means 'to be defeated.' I have assumed for the sake of argument that X u Shen intended the active meaning.  Note on Graphic Evolution Of the 26 bronze forms at NJWB 1352 (21 at OJWB 1096), only one is like the seal form in having the'person'separated from the'spear,'thus:  ( \§j | § >^JE ) .  So this is quite a late development  Usage (a) To attack (an enemy tribe) Verb:  (b) To decapitate (as a sacrificial act)  2  A type of sacrificial victim ('decapiturus') Noun:  1  The Guangyun only gives a r&sheng reading, but see Lu Deming (Pan Chonggui 1983.2:2020).  In the past, some scholars (e.g. Shang Chengzuo, ap. JWJ 8.2661; Shima 1958:335) argued that 4j£ referred to a kind of dance, since they could not believe (in spite of the archaeological evidence) that the Shang would be so barbaric as to perform human sacrifice on such a large scale. 2  106  The Shang perhaps performed gu6  'decapitation' on their enemies in battle, and this  would be how the word f& 4"x] 'strike' also came to acquire the specialized meaning 'decapitate.' It is this specialized meaning that the bone graph depicts.  JGWB 4388 (unidentified) S.465.1x2  Not in JWJ  Analysis The graph clearly shows a person poling a boat, but it is not known what later character it corresponds to.  This pictograph may have been replaced by a phonetic  compound character, but it is hard to determine which one. Professor Takashima has suggested cao > e  'to row,' and semantically this seems like a good candidate. The  meaning 'go by boat' fits well into the OBI context. The present graph is most likely a variant of J). S.465.1). This is one of the few examples of rin  ^  (JGWB 4389, unidentified;  and dji ^  alternating in the same  graph. It is not clear to me why the frontal view was felt to be appropriate in this variant.  Usage It probably means 'to go by boat.'  107  Section iii: Body Part  In a small number of graphs, an element is added to the 'person' element to indicate the part of the body where it is added. This element may be simply a circle, or it may be a phonetic hint, or it may actually depict the body part.  108  25 |  JGWB 4259 (unidentified)  S. 10.4x9  JWJ 8.2747: Jan j%  SW 8a.27a: jj  , { U -  SWDZ 8a.71b: Ii 7T person.  M  means xiaji ~p ^  Inji tfL JL  7\  H  . f|  : &  ^  M  ,  'base, support:' the buttocks are the base of a  is like saying zu6 yu chuang  In the variant j ^ - , sun Tr: Tun 7 ^  W  Ty* rf^- 'sitting on a board.'...  is phonetic.  'buttocks' means bj. '|f !^ 'haunches.' It consists of shi f*  [described at  SW 8a.26b as 'depicting a person lying down,' but I think it shows a person sitting], with ji 7T 'base' underneath situated on top of ji / L 'stool.' with r£y,  'meat' signific and sin % • . $  'bone' signific and dian  : iin  is also written is also written with g i  phonetic.  Analysis The graph consists of 'person' with a circle added to indicate the site of the fundament. JGWB also includes a form ^  (Yibian 5839, not in S), but the circle is on  the knee here, so perhaps this graph should rather be identified as xi Tsung-tung Chang (1970:110, n.l) proposes.  The seal form is somewhat more  complicated: the fundament has been distorted into (or replaced by?) j i (probably the original pictograph for ji ^  'knee,' as  T\  'base'  ) and a stool has been added underneath (rather  thoughtfully) for the person to rest his fundament on. The modern character is usually written  Usage  . Note also the derived qusheng word dian  'the rear (of an army).'  109  L i Xiaoding says that it is used only as a person's name, and I think this is probably correct, e.g.  (Bingbian 175.1) ...yin-day cracking, Dun tested: Tun will perhaps have illness.  This is unlikely to refer to an attack of piles, which one would expect in OBI to be phrased something like j L ^  $~ jfc  'the king has an afflicted fundament.' At S.447.2-449.4  there are no examples of 'body-part ^  4~  only 'person ^  f~  .'  Not in JGWB S.10.4xl  Not in JWJ  Analysis The graph shows a person with a circle behind the neck and a bar across the leg. The sole inscription reads:  (Kufang 283) As for the sick ^  , exorcize it to Ancestress Ji and Ancestress Geng.  110  So it is obviously a body part. By analogy with the preceding graph, we may safely conclude that the circle on the neck indicates the part of the body in question. One could tentatively suggest the translation 'neck.' The bar on the leg is a mystery. Perhaps it indicates qian T " . It is hard to say why it should incorporate this element. It could be a phonetic hint, though of course we do not know how the graph was pronounced. Note however that this inscription is from a traced collection, so its transcription may not be entirely accurate.  JGWB 1049: w i i S.27.4xl  Not in JWJ  SW Sb.la:  ,  flfc^.  JUL t\  4  K  ...  8? % K * f£ • Tr: Wei  £  f  fa  . -£ L  'tail' means j££i ' m £ 'slender.' It consists of mdo  [in the seal form, it is upside down] behind sju  &  fa  % j£, ,  'hair' upside down  ['body']. The ancients sometimes  sported an ornamental tail, and the barbarians of the south-west still do to this day.  JWJ overlooks this graph, which only occurs once, but it is unmistakeably wei , as JGWB has it. It is perhaps somewhat tendentious of me to include this graph in my section on 'human body parts.' As to why the creators of this script decided to depict the tail on a human being, one reason may have been that, if the tail had been depicted on an animal, since animals naturally have tails, it would have been hard to indicate that this was the focus of the graph, whereas when it is depicted hanging from a human being, it  Ill  immediately seizes ones attention, due to its unexpectedness, and thus there is no doubt where the focus of the graph lies. However, X u Shen's explanation gives us another insight. Compare the graph pj±  /  jljt (I.l.i.a.2645), where I concluded that the tail  indicates that the people the Shang used as domestic slaves were captured from some tribe which had the custom of wearing tails. The present graph occurs in the expression youwei ^  , which may perhaps be understood as youweizhe ^jf fk^  ^  'the tailed  ones.' The inscription reads as follows: 1  (Yibian 4293)  Jimao-day (16th) cracking, Dun ( J§ ?) tested: "Zhe will capture the Tu C M from Qiang (?)." The king prognosticated, saying: "Perhaps it will be bingxu-day (23rd). The capturing of the youwei will perhaps be on the xinday (i.e. xinsi, 18th?)." (Divined at) Bin (?).  Following a suggestion from Professor Takashima, I have understood Tu chu as meaning 'chu of the Tu tribe,' with chu perhaps referring to the soldiers or fighting men of that tribe. This fits in better with the 'capturing' context than Qu Wanli's interpretation in his kaoshi as 'foragers,' and supports the interpretation of youwei as a similar kind of alien tribe.  Although the common nominalization markers of classical Chinese zh& ~7& and su6 do not occur in OBI, Takashima (1984) has argued convincingly that unmarked nominalizations of a semantically comparable nature must be recognized in the inscriptional language. His proposals on this subject have enabled one to deal with many inscriptions that were previously hard to make sense of. The present case, where we have two verbs in succession, is a prime example, and we can make sense of this structure by interpreting the second verb as nominalized or as part of a nominal expression. 1  Shima's transcription (S.27.4) makes this graph look like y6ng $1 . The rubbing is hard to make out at this point, but in the corresponding duizhen inscription on this shell the graph appears more clearly as  2  . At any rate, in the present inscription it is a person's name.  112  2775 tH  Not in JGWB  S.420.1xl  Not in JWJ  Analysis The graph consists of 'person' and dp > which is the form of zhong *H / ^  that  occurs in ancestral titles in OBI. It is probably phonetic in the present graph. The context strongly suggests that it refers to a part of the body:  r£ * Y : E & f t * , *  . (Yibian 5405)  Wuwu-day cracking: Shi tripped and (caused-to-be-sick:) damaged her , it will not become an affliction.  M y interpretation of Shi as referring to a woman is based on a related inscription on the same shell which refers to her as Shi Fu TQ. (fp 'the Shi Lady.' The interpretation of the graph ^  as 'trip' is based on another inscription, Jinghua 1 (see S. 178.4), where the  context strongly supports this meaning. Graphic analysis provides corroboration, as the graph consists of  ^ 'steps' and ^  the idea of tripping down steps.  'person with foot emphasized,' thus representing  113  2838 >^*j  JGWB 4444 (unidentified)  S.434.3x2  Not in JWJ  Analysis The graph consists of qian j  I $1. (or ji  mouth open, and ^ * , which is perhaps dong $ sheng  / %. 'plant' element.  *b*r i  s  ?), representing a person with the /  on its side with an additional  probably a phonetic element, but its positioning  before the mouth could be intended to suggest that it refers to a body part in that area. The graph occurs after ni ^ ~ 'sick' in both its inscriptions (Cuibian 1266 and 1267), so there is no doubt that it refers to a body part, perhaps the throat.  2840  JGWB 4287 (unidentified)  S.434.4x2  Not in JWJ  Analysis This graph consists of 'person' with what appears to be dong  $ /  laid on its  side, and as in the previous example it probably acts as phonetic. The fact that the phonetic element is attached to the foot of the figure could also be significant, i.e. the graph could represent the word zhong {FJF 'heel,' for example.  Unfortunately the graph occurs  entirely without context, so there is no way of corroborating this, but this is the direction in which graphic analysis would lead us.  114  Section iv: Miscellaneous  In this section I give a couple of examples of graphs which do not fit into the preceding three categories. These are the graphs for niao Jjk liquid waste' and sJii %. 'solid waste.' If they are regarded primarily as verbs, then they could be classified under 'human actions,' but if they are regarded as nouns, then the \  element simply signifies  that the product is associated with, or conceived of primarily as being associated with, humans.  I include this miscellaneous section in order to show that not all graphs  containing \  as signific can be fitted into the three preceding categories.  115  9  ;{ . \i  JGWB 4262, 4263 (unidentified)  S.5.4x8  JWJ 8.2755: niao '%  ( jjL )  SW 8b. lb: / § , A  'V  4£  S W D Z 8b.3a: Often written  J&>. H JL , & &~ . in old books [this is the only form given in GSR,  1123d]. Tr: Niao ML 'urinate' refers to a person's small convenience. It consists of wgi M^, 'tail' and shui  water.  Analysis In the SW seal version the body has been corrupted, curiously enough, into wei 'tail', while the urine has been standardized as shui ^rC 'water.' However the bone 1  graph, as you can see, simply depicts a person urinating.  Usage To all intents and purposes it seems to be used in its original meaning, as part of various rites. However, the lack of context makes it hard to be sure of one's interpretation.  10  |  JGWB 1011: 4')  N  v  S.6.1xl6  JWJ 8.2751: shi  (not in SW), 4342 (unidentified)  116  SW lb.22a: g|  , £  j & . >U ¥"H , |  ' |  .  SWDZ lb.47b: In the Zuozhuan and S/u/i, this word is written with the character sju [whose primary meaning is 'arrow']. Tr: SM \% 'faeces' means f£n "ft ^  'manure.' It consists of cao V f 'grass' and wei  'stomach' abbreviated. The Yupian says that shi |fj  is commonly written shi yjjl .  Analysis As L i Xiaoding points out, SW overlooks the character shi  , only giving the  character shi |^) for this word, which is a semantic compound of 'grass' and 'stomach,' 1  suggesting plant matter being processed through the stomach. There is no need to regard the stomach as abbreviated, since H§ is probably the primary form of r6u l J A  '|^  3  , with the  'meat' element being added later as an auxiliary signific. Cf. wei Ja  , SW 9b. 16a, head form of wei * B 'hedgehog') contains IS) However, the character sM  (= hui  as phonetic.  must have existed in X u Shen's time, since it  existed in OBI, and has survived all the way down to the present. This shows that the SW is not entirely reliable for determining the antiquity of a given character, since there are indeed some important omissions, as various commentators have noted. The bone graph shows a person passing solid waste. In the modern character, the 'standing person' has been modified to shi f  1  , while the waste matter has been standardized as m£ %~ 'rice.'  Compare how the urine was standardized to shui Note also that shi  ^  Qi <  'water' in the preceding graph.  could also serve as phonetic (it differs from shi  furiously, however, SW (2b.4a) gives the seal graph /f^ This seal graph is clearly the character shi  as the 'old form' of si  , with the apparent huo  derived from the faecal dots of the bone graph and M have been made on phonetic grounds.  ci  'move towards.' element probably  'rice' added. The connection with xi  must  117  < *+8j'  only in tone), and indeed Karlgren puts sju  Jp;  in his sju f  phonetic series  (GSR 561d). Further evidence that the non-occurrence of sbi ^  in SW is merely an oversight  is provided by the fact that it occurs in the Shijing, though not in its original meaning but as a loan for a word xi 'to groan,' for which the SW form is  (2a. 13a). The Shijing line  (HY 66/254/5) reads:  "The people are now groaning."  Usage Used in its original meaning. In connection with tian ffl 'field,' it probably refers to muckspreading, e.g. 1  jkjH...i:%  » % tS ,ta * • + (Xucun 2.166)  Gengchen-day...tested: If next guiwei-day we manure the West Dan fields, we will receive abundant harvest. Thirteenth month.  Note on both the graphs in this section  In modern Chinese, niao Jp. niao  is used both as a verb and as a noun, whereas sju  is only used as a noun (the verb being expressed as lashi  ). The fact that  is a qiisheng word suggests that it is a derived form, and thus that the noun and  !See Hu Houxuan 1955.  118  the verb may originally have had different readings, though only the qusheng reading survives (as a noun, it has the alternative modern pingsheng reading sjii, but this reading is not found in the Guangyun  or the rhyme tables, and anyway does not seem to be  etymologically related to niao). The alternative writing with / | f , which otherwise stands for ni  nejk  'drown,' suggests that there may have been a rusheng reading. One would  expect the rusheng reading to be the verb, and the qusheng reading to be the noun. Since niao  now only has a qusheng reading, this is only a speculation, but the fact that it has  both nominal and verbal usages, whereas the non-qusheng  word shi JP>  has only a  nominal usage, could be seen as lending support to this speculation. In the above quoted inscription (Xucun 2.166), sjhi ip. is a noun used as a verb: it does not mean 'to pass solid waste' but 'to manure.' However, I do not think there is any need to try and force these two words into a verbal mould in order to account for the significance of the \  element. It is simpler to  recognize this element here as just indicating 'something to do with people.'  119  Section v: Phonetic  120  13 \  etc.  S.6.1x61 ( \  JGWB 273: gjan ^ x30, \  JWJ 3.721: ojan  ^  SW3a.3b: ^  T  ,  I  xl, \  ^  SWDZ 3a.6a: [Reads A  ^  x24, ^  • ^  +  'rin A  (for compounds, see 2182-2187)  x5, ^>  >^ l  xl*)  •  is phonetic.']  Tr: Oian ^r- 'thousand' means 'ten hundreds.' It consists of s b J -f-  'ten' and r i n J^.  'person.'  Analysis Karlgren (GSR 365a) says:  "The graph has 'man' with a stroke on one leg.  Explanation uncertain." However, the Duan Yucai text of SW analyses r i n  as  phonetic, and L i Xiaoding agrees with this. It is true that the initials seem somewhat disparate, but they are in the same O C rhyme (rin A *ts'g n), so it seems most likely that rin  ^  n i n < * n s j i , qian  4" ts'en <  is at least a phonetic hint. At the time that 2  the inventors of the Chinese script first came to write down the word for 'thousand,' rin A  was perhaps the nearest phonetic they could find. The bar across the leg is not 'ten,'  as SW has it, but 'one,' and indicates 'one numerical unit' (cf. the bar on top of Mi ~E 'hundred'). This bar came to be a permanent part of the character, but in OBI other numerals can be superimposed to indicate higher multiples. Although S only has erqian "T"  , sanqian ^~ \  and wuqian $~ 4" , JGWB also has one instance each of  This graph probably has nothing to do with qian -f" • I* occurs on Zhuihe 261 as the name of the north quarter. See Hu Houxuan 1956. Professor Pulleyblank has reminded me that the possibility of an *sn- cluster should be mentioned. Pulleyblank (1962:133) hyothesized the development *snh > *sth > *tsh, i.e. denasalization followed by 1  2  metathesis. Further support that he provides for this kind of development is the phonetic role of £i — ni  h  in c l 'jlZ t s ' i and Tibeto-Burman words for'seven' (Chinese gi "t t s ' i t ) with initial sn-. h  121  siqian rrp ^ -fc_  > liuqian T \ ^f-  and jiuqian -f\i  4"  and baqian J V 4"  , and it is quite likely that qiqian  ^~ also existed. Compare the multiples of bai "§" 'hundred'  (JGWB 485, 2175-2181). Even in OBI the higher multiples are sometimes written analytically. S records sanqian  =  ^  (Qianbian 7.24.2), wuqian ^ - £  1.31.5. mistakenly recorded in S as 1.31.6). and erqian  =  £  (Houbian  (Liulu.Shu 81). These  analytical forms are however quite rare.  Usage Mostly in deng ren  inscriptions, i.e. conscripting so many thousands of  people for military purposes.  1282  ^  JGWB 876: ruin i f  S. 194.2x530 (excluding  JWJ 7.2365: niin  SW7a.l8a: ^  Tr: Nian i f  ^  ^  )  if  , f£  %\  U  , 4  #  . &  means 'the grain is ripe.' It consists of h £  41  #  &  :  £  ^  'cereal' signific and qian  phonetic. The Chunqiu says: "There was a bumper harvest."  1  Analysis  207/Xuan 16/4 Jing. Legge (1861.5.1:330) translates: "There was a very plentiful year." I have given my own translation in order to harmonize it with the SW definition.  122  A  Ren S W, ren A  , qian 4"  a*"*  m  a  " if-  is phonetic in qian -f"  phonetic in injn if- . In OBI, njin if-  are all in O C zhen )@_  rhyme. According to  (in the Duan Yucai edition), and qian T simply has r i n  is  phonetic. There are no  examples with qian "f" , until the bronze script, where we find that out of 141 examples at NJWB 1164, approximately 28 (some forms are ambiguous) have qian -f" 1  (  m  12 of  these instances a further line is added at the bottom, so that the lower element actually looks like ting -i- ). From the point of view of the initials, rin nian 5f  than qian T  is. so the replacement of ren  is clearly much closer to phonetic by qian T  phonetic  is mysterious. L i Xiaoding explains the bronze form of qian "T" , along with that of r i n i  ( I  > 5 ), by saying that 'in ancient characters a horizontal stroke is often added,  but it is of no significance' ( ^  >C %  4r\  %.  )• He appears to  be saying that the apparent qian -f" element in nian 5^- is not really qian -f" , but simply rin  A  with a stroke through it (i.e. it evolved into qian -f" unintentionally, and  the further evolution into ting JL. is certainly hard to explain any other way). I have an idea that perhaps the bar applies, not to the rin indicating 'one' year, just as in qian \  and b j i  alone, but to the whole character, i.e. I^J the bar indicates 'one numerical  unit.' The coincidence with qian 4~ would thus be unintentional. The hi x^L element is always written on top of the rin  A - element, so Ye  Yusen's suggestion that the graph depicts a person carrying the harvest on his back probably has some validity. The suggestive use of phonetic elements is very widespread in the Shang script.  Usage  i Q J W B 941 has 117 examples, of which about 20 have qian  4"  123  In OBI, nian jifwhich §1  is nearly always used to mean 'harvest,' hardly ever 'year' (for  'ritual cycle' and sui ^  1  'Jovian period' are used), but in Zhou bronzes it 2  is commonly used for 'year.'  ^ o r the few examples, see Hu Houxuan 1942 (rp. in 1944) and 1987. See Pankenier 1983.  2  124  Part 2  The graph shen ^  / %  depicts the human body with the abdomen  emphasized, and occurs in a few graphs denoting words that have some connection with this part of the body.  125  16 ^  JGWB 3345 (unidentified)  S.6.4x13  JWJ 8.2719: shen  SW8a.l8a:  ,  &  . fa k.  *~  SWDZ 8a.47b: The phonetic is really shen  & ty  . &  ^  ,  *%  1  abbreviated. Y i /  f  . K  ...  cannot be phonetic  because it is in the sixteenth rhyme category [i.e. ji *T* ] while shen -TT is in the twelfth [i.e. zhen IJ|  ].  Tr: Shen Jj' means gong  'body.' It depicts a person's body. It consists of r i n  'person' signific and yi  )  phonetic.  The quibble over whether shen  ^  contains y i f  or shen  ty  phonetic is  irrelevant since the character is simply a pictograph in origin, and I am surprised that Duan did not realise this.  1  Analysis L i Xiaoding is of the opinion that the present graph depicts a pregnant person. But really,  Y is like jian % I iL  , w i n *f  point emphasized. Compare yun ^  / ^  / 1*9  , ai  i  /  (no. 19 in this Part) which shows someone  who really is pregnant. 'Body' is the usual meaning of shen ^ shen ~^r\  etc., in having the focal  , and the expression you  'have body' is probably just an idiom in origin to express the idea of  pregnancy. Thus Li's surmise that the OBI expression ni shen ^he seal form of shen  !J3  is  ^  , the 'old form' is  ^  refers to  , and the zhduwen form is  ^  (SW  14b.15a). Professor Takashima has pointed out to me that shen and shen ty are homophonous (both pin<*+ep), and that many pictographs do contain endomorphic phonetics. Thus although the bone form of shen  Jj'  does not contain the bone form of shen  £^ , which is ^  (see JGWB 1708), it is  quite possible that the graph developed in such a way that the seal form came to have shen incorporated as an endomorphic phonetic.  ty  126  complications in pregnancy must be dismissed, especially in view of the fact that it is used of males (indeed, there are no definite references to females):  •fct  (Yibian 7797) Tested: The king's sick body is due to Ancestress Ji's curse.  Usage In its original meaning 'body.'  JGWB 3599 (unidentified) S.6.4x2  Not in JWJ  Analysis The graph consists of shen examples occur in the expression  'body' and ydu t  ^  S.7.1xl  'hand.*  Both  , so it would seem to be a variant of shen  , though the addition of the hand is mysterious.  19  / yL  JGWB 1695: yjjn %  127  JWJ 14.4315: yun  SW 14b.l2a: ^  ^  ] £ J!L  ,  . M  S W D Z 14b.24b: [Emends c6ng ji IX J L  £  , AA / b .  to nai sheng  ft  ^  'nil $  i s  phonetic.'] Tr: Yun ^  'pregnant' means 'to contain a child within one.' It consists of z£ ?r  'child'and j i  Jb  'stool.'  Analysis The graph shows a woman bearing a child in her womb (a sort of X-ray picture), thus illustrating the meaning 'pregnant.' L i Xiaoding readily recognizes that the bone graph does not contain nji  , and yet he approves of Duan's emendation to SW making  this the phonetic. Karlgren also recognizes nM accurate, one should say that the original shen ^ perhaps replaced by?) nai X5  as phonetic (GSR 945j). To be element was either corrupted to (or  as a phonetization.  Usage It is used in its original meaning, in the sole inscription:  <kh. &  S mistakenly has this number as 584.  (Yicun 586 ) 1  128  Yihai-day cracking, Dui tested. The king said: "[She] is pregnant, it will be good (i.e. a boy)." Fu said: "It will be good." 1  20  ^  Not in JGWB  S.7.1x3  Not in JWJ  Analysis The graph is a combination of shen \  I  'body' and n j . ^  /  'female.' Sometimes the addition of the female element does not seem to alter the meaning I ' \  (III. L i . 1859  and S.275.1). Thus the present graph is probably the same word as shen  . In the  of a graph, e.g. mi£ ^  case of mie ^  ~ ^  / ^  (S.212.1) and bjn $1  , the reason for the variation is that the ^  ~ ^  element is probably  phonetic, and thus has no effect on the significance of the graph. ^  and ^  also  exist as independent graphs (S. 107.2), and there does seem to be a difference of sex here I;  (see I.l.i.b.657). In the case of bjn -j^ , Professor Takashima has suggested to me that the kneeling figure <k represents a human being regardless of sex.  Thus there is no  conflict of gender in this variation, only a difference of whether the gender is specifically marked as female or not  Usage iThis implication is deduced from Bingbian 247.1/2, in which the vanci Jff^L in this duhhen say " ^ Keightley 1978, fig.12.  jfe  , %-  of both divinations  " 'it was not good, it was a girl.' For a full translation, see  129  In all three occurrences it refers to a woman, Lady Hao ^ preceded by ni $~  4&  , and is not  'sick,' so it could perhaps be being used in the specialized sense  'pregnant,' e.g.  fi * h > it i : k  § %> &  <f  • (Bingbian 340.3)  Bingshen-day cracking, Que tested: Lady Hao's pregnancy will not result in the Lady dying.  486  |^  .jjjj^  JGWB 3985,4324 (both unidentified)  S.87.3x2  JWJ 4.1509: f i JWJ 8.2672 (second form only): \%^ (not in SW), same as  S W 4 b . 9 b : f l , J f . MJ . JUL A S W D Z 4b.25b: H6u / f  , £. ^  5^  -a  HI  .  is a rhyming [i.e. paronomastic] gloss. This is just like other  SW definitions such as "& "§£ 'hair' (*pat) means M HC 'to pluck* (*brat)" and "wei 'tail* ( * m a P ) means wei ^>C 'fine' (*mal)." The gist of the present definition is that the abdomen was named for its thickness. The Shiming[.Shi xingti] says: "Fu Hg. 'abdomen' (*pek ) means ft <f|_ w  'doubled' ( * p 9 k ) or fu. § w  'ample'(*p9ks)."  This  is the same mode of exegesis. The [Erya.]Shigu[HY 3/1B/30] and the Mao commentary [on the Shijing] both say: "Fu  means h6u )%  'thick' (*gaw?)." [I have translated  130  Duan's commentary at some length, as it throws some interesting light on the nature of Chinese paronomastic glossing, or 'punning definitions.'] Tr: Fy. andfji  'abdomen' means h6u )%. 'thick.' It consists of r£u_  'flesh' signific  j | _ phonetic.  Analysis The first graph consists of shen % A  md fu. fii | L and  and fji ^  , while the second consists of ren  . They may be transcribed as * |JL  and *  respectively (neither  character is in SW).  Usage The first is used to mean 'abdomen,' while the second is used to mean 'return' and may thus be regarded as a variant of fji ¥  /  (\  ), as Qu Wanli correctly states. It  seems possible that the difference in signific indicates a difference in word denoted, indicating 'connection with abdomen' and h  \  indicating 'human action.'  (Xubian 5.6.1) Guiyou-day cracking, Zheng tested: The king's abdomen is uneasy, it will not be prolonged.  (Jiabian 587) Do not return.  1  iThis is the interpretation suggested in Qu Wanli's kaoshi 587.1.  131  -  Part 3 :\  Although in the bone script we find the upside down figures ^  !(see II.2) and the kneeling figure  \  \  (see 1.6) and  (see III), there is, strangely enough, no  recumbent figure * - i r - . In graphs where one might expect such a figure, such as those representing words to do with sleeping, dreaming, sickness and death, one finds instead \  . That  \  here is intended to represent a figure lying down, rather than simply a  person as a participant without any implication of posture, is something that I infer from the orientation of the other elements with which it is combined. We find it combined with the things that the person is lying on, to wit 1-r'a mat,' and £  'a wellhead-shaped mortuary frame,' H  'a bed.' These are objects which in real life are usually horizontal, but 2  in the bone script the signs for them are written vertically, as part of an orthographic convention which we also see manifested in most animal graphs (e.g. sM I  'dog' ). Note however that the mat does occur horizontally in  the unidentified graph  (IH. l.i.2028), which shows a person kneeling on a mat. This  'pig' and quan y  3  suggests that when the mat is written H  , it is put on end as an orthographic convention.  though these are admittedly rare. 2  The identity of these three elements is explained later in this Part.  3  But note that lu  drawn like nul P*^ /  'deer' is not treated in this way, probably due to the fact that the head is @  'eye,' which is usually written in this position, and maybe it would have  been less convenient to draw the eye and antlers sideways.  As for nM  iji  /  ^  'ox' and yang  'sheep,' these are frontal views of the head only with their distinctive horns schematized.  132  As independent graphs, £t  ,  H  and  (1  only occur in the vertical form.  perhaps the fact that these graphs are already vertically oriented that led to the use of  1  It is \  in combination with them rather than the creation of a special recumbent form. Actually, the graph  could perhaps be seen as an aerial view, and the variant ®  support this, but H  and  $  would  cannot be regarded in this fashion. Another indication  that the person is intended to be understood as lying on the mat or the bed is the fact that the person element is always drawn with the back to them. Thus, although the graphs as a whole may be left or right oriented, we never find *  i t -11  , only  andlf -JB .  ^ee S.301.1, 447.1 and 412.3 respectively. However, in the graph does not depict a wellhead, but a wellhead-shaped mortuary frame, so perhaps it should not be identified with the independent graph jing ^4 / "H" , which does actually differ slightly in that the lines are somewhat concave, whereas the lines of the mortuary frame are always straight.  133  62 m  JGWB 1015:  .lit.  S.12.4xl52  (not in SW)  JGWB 784: ffl , [ £ ] =gjji  ~M  JWJ 4 . 1 4 5 3 : ^  .fit  - IS  ~ IS  0  ^SI ? L  SW 4b.6a:  ^ .... Hi  : £ £ ?L -4« j i b  SWDZ 4b. 13b: Si Tr: S i  is defined in Fangyan as jin ^  'to die' means  behind. It consists of 6 SI  is like this.  .  si *|tjj" 'to come $  'to come to an end.'  2  to an end.' It is that which a person leaves  'remains' and r i n A_  'person.' $  : the old form of  3  X u Shen uses the word  si  sfcfc  rather loosely for the sake of getting a  paronomastic definition. The second definition (if I have interpreted it correctly) appears to take si  ?t  as shi SL  'corpse.'  Analysis Before analysing the present graph, it is necessary to discuss its identification, for although it is now generally accepted as si (of the sort that we find between \tl  , there seems to be a graphic discontinuity  and you  form as the non-existent character  , saying that 'it depicts a person inside a well,'  and identifies some of the variant forms as cuji Q ijGWB 783 has the last two graphs as yjn. IS ^Fangyan Fangyan  lla.21b defines si that si 3  'prisoner.' However, the OBI context  • They occur in S at 40.1.  3/24/49. The text goes on to explain jin 13/87/135 defines si  ). JGWB identifies the principal  as s M  as jin sheng  . which Guo Pu's note glosses as  as  'to finish life.' . SW  'water is exhausted,' i.e. dries up. I can find no evidence  could mean anything like 'corpse.'  This old form occurs in the old form of yi  ^  ft  (see I.l.i.b.2385).  134  suggests that the graph means 'die.' It is hard to find iron-cast proof, but the following inscription is perhaps as near as one can get:  s * t v i. • I 0 3» c & i a  . (Jimbun 446)  Bingwu-day [cracking, X] tested: [X] will recover from his/her illness and 1  will not die.  This is the only inscription I have been able to find where the context more or less demands the interpretation 'die,' and it is unfortunately rather fragmentary. However, the filling in of the missing parts in the above transcription, which is that of ltd Michiharu (the transcriber of this collection), is well justified by numerous parallel inscriptions dealing with sickness. For a thorough examination of the merely suggestive inscriptions, see Hu Houxuan 1970. Only the first form given by L i Xiaoding is the graphic antecedent of the modern character si  . Strictly speaking, although the present graph probably means 'die,'  there is no proof that it represents the same word as si  ft.  . It could represent another  word with the same meaning. This may be being too cautious, but the point should be borne in mind. Ye Yusen's objections to identifiying the present graph as si  ft  are that dead  people do not stand up in their coffins, and that the person's head and feet are sticking out (assuming that the graph is intended to depict a dead person lying in a coffin). As to the first objection, I think the present section of my thesis should leave no doubt that  ^  must be recognized as sometimes representing a person lying down. As for the second objection, archaeological excavation has revealed that, in the late Shang medium size tombs at Anyang, the boards that make up the outer coffins (guo !For my interpretation of this common formula, see Takashima 1980.  ) intersect near the ends to  135  form a shape like the Chinese character jing  'well' which actually depicts a cross-  frame on top of a well. Not only is this the shape seen in the present bone graph, but the term jingguo  .  4^  A  #  'well-frame coffin" also occurs in the Yili:  ¥>) >fe *  y.A *  #9 f  ft  (Shang Zhou kaogu, p. 101)  The coffin chambers were all more or less oblong. In the tomb at Hougang, M32, traces of the wooden outer coffin were quite well preserved, and consisted of timbers forming a pattern like the character jing $~  , just like  the so-called 'well-frame coffins' often found in Zhou tombs. The Yili. Shisangli  says:  Jf%  "having framed the coffin."  1  The  commentary to this reads: "When the carptenter makes the outer coffin, he cuts the wood, and constructs it outside the funeral booth in a shape like the character jing  ." The Yili zhengyi quotes a scholar named Chu  Yinliang as saying: '"Well-frame construction' means that the beams of the outer coffin are placed two lengthways and two breadthways, piling them up alternately and successively, in the shape of the character jing  In some of the small size tombs, it was also found that the guan  ."  'inner coffin' had  been supported by two transverse lengths of wood, which happened to be directly under the shoulders and calves of the occupant. The inner coffin having rotted away, this now Steele (1917.11:74, % 18a) translates "When the outer coffin has been brought to the door," as if jing 4$  meant 'bring to the door.' I do not see how he gets this interpretation.  136  gives the spit and image of the bone graph  , and this is perhaps what the graph is  intended to represent, rather than a person inside a coffin. A burial of the Warring States period excavated at Huairoucheng  in the Peking area exhibits  exactly me same thing. For convenience I reproduce both below: 1  (Shang Zhou kaogu, p.105, fig. 80)  In the variant  I  2  (Zheng Liangshu 1971:234, fig. 48)  , the dots perhaps represent earth thrown onto the coffin.  In view of the above archaeological information, I shall describe the !Jlf" element in the present graph as a 'mortuary frame.' I do not know what its function was, but since it is depicted in the graph for 'die' it was clearly a conspicuous part of the funeral paraphernalia.  Usage Always used in the meaning 'die.'  Though curiously the author here makes no comment on the crossbars beneath the skeleton's shoulders and calves. The photograph on which this drawing is based may be found at Ma Dezhi 1955, Plate 1. 1  2  137  2026  {I  JGWB 848: si  ,4322 (unidentified)  S.301.1x6 2027  JGWB 907 si A%  S.301.1xl0  JWJ 7.2436: (both forms) si  SW7b.5b:  (fa,  Tr: Su >fi§  . M ^ , >ffi f .  ±  means zhi  phonetic,  'to stay.' It consists of mian  ,  ^  * L  *  .  'roof signific and sji  is the old form of si  Analysis Graph 2026 shows a person lying on a mat. L i Xiaoding (JWJ 3.689) identifies the mat with the SW (3a.2a) character tian  1  t§  , and says it is the old form of dian  2  (SW 5a.3b) 'bamboo mat.' Graph 2027 has a roof added, thus underlining the idea of 'passing a night' in a building. SW maintains that  is the old form of si  .  This would seem to be a guess based on the fact that they are homophonous, but it is undoubtedly correct to treat follows SW, and also includes  as a variant of si tf& , as L i Xiaoding does. JGWB ^}  (my III.l.i.2028) but, as L i Xiaoding notes, this  shows a person sitting, not lying, on a mat, and its OBI context is also different.  Usage  ^he Guangyun gives this character two readings, -bfl 2  t'em >tiia.  Guangyun  h  f\JL  5& -£77 dem'.  ^fjfe^L.  tV  t ' 3 m > t i n and h  4$L>  138  Probably in its original meaning 'pass the night,' but the context is rather limited.  2917  JGWB 964: ni  j~  S.447.2x435  JWJ 7.2515: ni  $~  SW7b.lla: | " , J%  <H  . X  I  ^  ^  .  ft  $f  %  ift  •  ...  S W D Z 7b.26a: [Suggests that y i reading of ru $~  ff*r /  9  i a ' < * al'? is a paronomastic gloss, but if the ?  nre j k < *nrak (?) is correct, this seems rather strange. The reading r  of this character is a problem. GSR 1260b notes that the Qieyun takes it to be a variant of chuang  A] means'to recline.' It refers to human sickness. The character represents the  act of reclining.  Analysis The graph shows a person lying on a bed sweating, thus illustrating the idea of sickness. Some people identify this graph as jf ^ identification as ni $ ~  . L i Xiaoding himself accepts the  on graphic grounds, but maintains that it is the primary form of jf  !Wang Renxu (Long Yuchun 1968) has this character under yang jf^ and the definition bing  rhyme with the reading "i"  'sick,' and notes that it has the alternative reading - ^ C  » but, as Long Yuchun (1968:212) notes, the character is not listed again under mii rhyme.  ^  139  •^tj^ . with shi ^  being added later as a clarifying phonetic . The phonetic value of 1  the textually unattested character ni f  is a problem that is hard to solve.  There is little doubt that the seal graph \ via the bronze element Ijf  evolved from the bone graph \ \  (see O/NJWB 1030-1033/1270-1276). As Yang Shuda  correctly remarks, the horizontal bar in H J  is a vestige of the person who lay on the bed  in the bone form. Yang also notes that SW (7b. 11a) gives ^  as the old form of j j  , and this form preserves the bronze stage of development of the person lying on the bed. Note also the following bronze graph, which occurs as a person's name, but which from its components would appear to be identifiable as ji ^  :  (NJWB Fulu 2.348; not in OJWB between Fulu 2.27b.2-3)  My decision to follow the identification of  [j "j" as ni  ^ ~ rather than as ji  based on purely graphic grounds: the bone graph contains no shi If we accept L i Xiaoding's claim that ni f  'arrow' element.  is the primary form of jf  7 ^ , then the  Qieyun reading of the former must be wrong. However, SW does not say that ni the old form of ji  In its original meaning 'sick.'  lT  t seems likely that sfii  ^ ~ is  , and I have decided to go no further than the evidence allows.  U s a g e  2922  is thus  JGWB 5859 (unidentified)  <  ? is intended as phonetic in j i 0^  dzit < *dz9C  140  JGWB 4218 (unidentified) S.450.1x2  JWJ 7.2491: (first form only)  (notinSW)  Analysis The present graph consists of 'person and 'bed' under 'roof.' The first form given 1  by Shima also has a kou  E  'mouth' element.  Under the bronze graph flf}  is  (O/NJWB 1025/1265), Rong Geng says:  Ail.  % M jtt.^ f SW has meng  1  radical but no ^  radical. The present bronze graph  . It is the radical in characters such as yyji jf£  'to wake' and mei  'to go to bed.' This is Gao Jingcheng's theory. 2  There is only one example of this bronze graph, on the vessel  ^  JjT  , and  unfortunately it is accompanied by only one other graph, so there is no context to corroborate Rong Geng's identification. However, it makes good sense to set up jjp  as  the radical in the group of characters at SW 7b. 10a that X u Shen analyses as containing abbreviations of his meng ^  radical, and I think that the present bone graph may also be  identified with it. It shows a person lying on a bed under a roof. The SW characters containing this radical all have to do with lying down and/or sleeping. analyses the present bone graph as containing ni regards as the primary form of j i  , which, as already mentioned, he  'sick.' However, it should not be analysed as a  iThis graph also occurs in the bones, as a place name (S.447.1, JWJ 7.2489, JGWB 934).). I have been unable to trace the identity of this person.  2  L i Xiaoding  141  sick person lying down, but simply as a person lying down under a roof. That is to say, [j^J means something different from !j { .  Jj ^ I ^~ occurs in characters to do with  sickness, whereas (f^j / j p occurs in characters to do with lying down and sleeping. X u Shen tries to work the meaning 'recline' into his definition of ni ^  in order to get a  paronomastic gloss (supposedly), but only the second part of his definition 'human sickness' really defines the meaning of ni  , as can be seen from the fact that all the  characters in SW containing ni ^ ~ signific refer to sickness (none simply refer to lying down), while all the characters containing  refer to sleeping (none to sickness).  also has dots added to show the sweating, whereas f\§\ no sweat involved. Compare s i  does not, because there is  / rf§ , which shows a person lying on a mat under  a roof. There is no sickness involved here, and no sweat is portrayed.  Usage Unfortunately the usage of both graphs is rather obscure.  2925 j} R •  0  JGWB 4166 (unidentified)  S.450.2x50  JWJ 7.2527: y_6u. 9&-  SW7b.l2b:^  , tf  • M  t  .  S W D Z 7b.29a: This is the same word as y6y. JC\ Tr: Y Q U ^  means zhin *f f  • [SW 9a.5a, same definition].  [defined at SW 9a.5b as  not right']. It consists of the sickness radical ni 7  *f  and y6u  ^  jE. phonetic.  'head  142  Analysis The graph is to be analysed as in SW. It cannot be analysed as * ^  and  fj  , because  go together as a unit. To transcribe it as xi znl is even more out of the  question. L i Xiaoding also maintains that y6u which SW 7b. 12b defines as xiao fu bing  5^  is the same word as zhou ^  /]>  ,  'a sickness of the small  stomach,' and that this explains the variant ^ fj in which the stomach is drawn in. He says that the 'hand' element is not only phonetic, but also represents the idea of stroking the sick stomach to soothe it. However, you  and zhou  have quite different  meanings (if one accepts their SW definitions), and they are phonetically too far apart, having both different initials and finals in OC.  Usage Mostly a person's name, but could refer to a sickness in the following inscriptions:  (Yibian 2340) Tested: Sick body ( 7 ) improve. 1  m Jg. K  . m.  i  : ' f -fc  A &% . {Jiabian 2040)  Bingchen-day cracking, Que tested: Lady Hao's sickness will continue to improve.  In view of the uncertainties, I have refrained from being specific about what sort of sickness it was. However, the present graph is a good example of the heuristic application of the present theory of graphic analysis. Even though there is some dispute about the ^ee 1.2.17 for the interpretation of this graph.  143  identification of the present graph, we are able to predict from the human figure element lying on a bed that it will probably have something to do with sickness, and this is corroborated by those inscriptions in which it is used in its primary meaning (as indeed one can often guess the general meaning of modern characters one does not know from their radicals). Thus although it is nearly always used as a person's name, we can be quite confident in regarding this as a loan usage.  JGWB 962: meng  2926 S.450.3xl47 2927  .^'B  JGWB 861: rt)ru (not in SW)  S.451.2xl3 2929  .  fP  S.451.3xl2 2930  JGWB 962: ming J j ^  S.451.3x7  JWJ 7.2509:  SW7b.l0a:|>wt , jtf ffl  0  M % A> .  ...  %% ik 8 H E m. £ ^ $ 2_ £ &  4k *  >x  ^ tl  . ^  ...  SWDZ 7b.24b: Now written ^ phonetic and signific.  . ^  means burning >JN  8$  'dark,' so it is both  144  Tr: Meng  jffi  'roof,' ni  i  'dream' means 'to sleep and have sensations.' It consists of mi an 'recline,' and meng  rf  1  ^  is phonetic. The Zhouli [says:] "Par les  positions du soleil, de la lune, des planetes, il devine les presages heureux ou malheureux des six sortes de songes. Ces six sortes de songes sont:  lo les songes rdguliers; 2o les  songes terribles; 3o les songes de rdflexion; 4o les songes de veille; 5o les beaux songes; 6 0 les songes de crainte."  2  Analysis The OBI context shows that graph 2926 *j ^ meng  /  must be a variant of graph 2930  'dream.' It is by far the commonest form, and L i Xiaoding says that it  is an abbreviation of the latter, the head part being abbreviated and distorted. However, it is a rather curious abbreviation. ^ hair ^ 7=L  is probably the original form, and the dishevelled  suggests a person in the throes of a nightmare.  (^nk*. )  3  Compare wei ^  /  (example from Fuyin.za 58), which shows the hair of the head all pointing  in the same direction, and mgi  ft*/  M  (example from Yinxu 1854), which shows the  hairs of the eyebrow all pointing in the same direction.  |^  may thus be interpreted as  showing the hair pointing in different directions, i.e. dishevelled. Compare also dou 4  (example from Yibian 7119), which shows two people fighting. In this  example, the hair of the person on the left is particularly wild.  In ru6 %r?l I  (example from Jiabian 205), the hair seems to be smoother and more orderly. The variant 2927  appears to contain hjL ^  explanation for this at present.  *As we have already seen, SW defines ni  The variant 2929 ^ffJ  ^  / Jfi, 'tiger'-I have no appears to contain jian  both as 'recline' and 'sick, but analysis of its use as  signific, compared to that of n , shows that it only means 'sick.' However, I have glossed it as 'recline' here, because I feel that this is the meaning that Xu Shen intended. The meaning 'sick' does not seem to be so relevant to dreaming. H Y 6/28b. Translation from Biot 1851.11.82 (XXIV.27). See 1.7.36. See 1.4.98.  2  3  4  145  ^  I Wi 'see'-this can be interpreted as representing the fact that in dreaming one  seems to see things: visions float before one's inner eye. The final variant, 2930  ,  has the eye elaborated by a brow, and it is from this form that the modern character develops. The eyebrow, m6i ^@ , is perhaps intended as a phonetic hint, since it has the same initial as meng ^  / ^  ^  muwn. <*m9n.s.  This element ^  h  (S.212.1) met < *m at. where it probably stands for m& j  phonetic. In view of the phonetic value of m6i , which he transcribes as #J  is also found in mie "l  mi<*majas  % , Sun Haibo (ap. JWJ) suggests that  , should be identified with the SW character mi  , which is defined at SW 7b. 10b as  'not enough sleep.'  However, the meaning in the OBI context has to be 'dream.' The reason why m6i ^@ not a very accurate phonetic in meng ^  is  is perhaps that it was not planned as the  phonetic in this character, but came about through a modification of what was originally a signific element, i.e. it is a phonetization. Note however that  ^  appears to be  &  &  phonetic in characters like meng rp* , meng /g"* , m6ng ^ & y£j  . This is perhaps by analogy with the character meng  -&£  Usage In its original meaning 'dream.'  2931  f)^  JGWB 2101: meng a  S.451.3x3  JWJ 7.2509: ming a  Analysis  %  •  %  #L  , mang /j-f^ and hong .  146  JGWB and JWJ both analyse the present graph as a h6w6n ^ and fu.  >C  of meng  ;C_ . This would seem to imply that it is an ancestral title, 'the Meng  Father,' whoever that may be. Ding Shan (JWJ 7.2513) identifies him as the sage Fu Yue »TL  > whom Wu Ding was led to by a dream and thus might have been given the  title 'Dream Father'-an interesting suggestion, but impossible in the OBI context (Ding makes the context seem more appropriate by misinterpreting you |L  and shai  /  1  as mji  ^<  I ^  \jd  as zhi  X.  ). In the context it seems best to interpret it as some  kind of oneiric omen, e.g. a nightmare.  It may be transcribed * ^fc .  The fullest  inscription in which it occurs is the following:  Oracle bone specialists, following Sun Yirang (JWJ 9.2997), usually transcribe this graph as which SW(9b.l5b) defines as'a long-haired beast* ( in the Henei area, and says is read like di  ^  JIfc  ^  ^>  ) or the word for'pig'(sM  (the Guangyun actually gives two readings:  , )  ^ \  "fcfl dej > d i and % ~§~ j i > yj). If this is correct, then the OBI usage must be a rebus for something. It occurs both as a noun and a verb, and contextually the translation 'a curse, to curse' seems to make good sense. Although the graph does not look very animal like, one may compare it to qM h  h  'to seek,' the primary from of gnu. 'fur clothing,' probably depicting a pelt (JWJ 8.2733). (The theory of graphic analysis used in this thesis would lead us to look for a common symbolism for the  ^  element in shai  ^  and gju ^  . If one can recognize the same element  in wei  , then perhaps it is the stylized drawing of a tail.) Guo Moruo (JWJ 9.2998) accepts  this identification, but also notes that the graph is very similar to the graph gives as one of the 'old forms' of sha~shai bronze form of the phonetically similar cM the  TjO element of gha.  ?e t~§e j r  r  h  ^  which SW (3b. 13b)  < * s r a t ~ s r a t s , and also to ^  ts'aj <*ts'ats. I think ^ h  , which in the seal from is written  ff  v  , the  may have evolved into  , but the meaning 'kill' does not  seem appropriate to the OBI context, so I read it sak, in which reading Lu Deming defines it as hM 'to harm' (see Pan Chonggui 1983.2:1606). The OBI usage should perhaps be understood as 'harm, to harm.' However, I still feel that there are some problems with the identification of this graph which would require a full separate study.  147  % *\ H 1 • ¥ $ fc. * *. fir . (Tongiuan 430 = Jinghua 6) Guichou-day (50th) cracking, Zheng tested: "In the next ten days there will be no misfortune." The king prognosticated, saying: "There is a curse, (I?) had a nightmare (?)." On jiayin-day (51st) there was indeed a coming 2  trouble-report.  Zuo reported, saying: "There were Tu chu from Y i ,  3  (numbering) twelve persons."  Since both meng ^  and fu.  have phonetic series (GSR 902 and 102), it is not  certain in the present graph which is signific and which is phonetic. However, X u Shen did not set up fu A . as a radical (he classifies it under y6u  at SW 3b.8b), and the  Kangxi dictionary, though setting it up as a radical, is only able to offer under it colloquial terms for 'father,' plus a few rare characters, so perhaps fu.  was not used as a  signific till after the Han dynasty, though the lack of dictionaries for Han and earlier times makes this conclusion tentative. The Fangyan does not have any characters with fu as signific, though there must have been different expressions for 'father' in different dialects at that time. It seems most reasonable to assume then that in the present bone graph, the meng  element is signific and the fu.  element is phonetic. Guo  Moruo makes the same assumption, in his commentary on the above cited inscription, when he suggests that it is a complex form of fu ^  The identity of X  lr  as Hi  . SW (7b. 11a) simply defines this  , rather than as a variant of wang  , is supported by the  writing ^ on Yibian 4293 (quoted under I.l.iii.132) I feel that a past tense translation makes best sense here, although of course in the original bone text there is no tense marking. 2  3  This translation of jian "^T is a functional one based on its OBI context.  148  character as blng  'sick,' and quotes the Shijing line  4fc  $JL 7 ^  driver is ill" (HY 1/3/4). However, I have already shown that ni {|!^ / ^ |j ^  - ^ L "My and meng  as signifies are generally quite distinct, and one would not expect meng  ^  to be  used in a character denoting sickness. Furthermore, the OBI context suggests that the present graph refers to some kind of bad omen, as do shai  &  and jian "^T , so another  suggestion one might make is that it represents the word bji 11$ 'a malevolent and noxious deity' (GSR 102k'), which occurs in the Zhouli.Diguan.Zushi passage it  #  iff  fc-h  Z.  ^  "Au printemps et en automne, quand il sacrifie aux  esprits malfaisants, il fait encore de meme." Zhu Junsheng (1834:9.29a) says that this is 1  the same word that is written M "Ty in the ZhouliXiaguanJiaoren passage ^j?  ^  ?f>  "En hiver, il sacrifie au mauvais genie des chevaux." In his commentary on 2  this passage, Zheng Xuan describes mabu  % ^  as  ^  ^  j§  'a spirit that harms horses.' Although it is impossible to say for sure that the present bone graph denotes the same word, the meaning of M  l l $ ~ "fp~ is certainly apposite to  the context. Since the Zhouli writers did not have a proper character for this word, writing it with two different phonetic loans, one might suggest that the proper character was in fact *  , but that knowledge of this character had been lost.  2932  -fHr  Not in JGWB  S.451.3xl  Not in JWJ  iTransladon from Biot 1851.1:254 %26. translation from Biot 1851.2:257 1)48.  149  sw7b.iob: $[ , % HP £ /fc . M f S W D Z 7b.25b: [Emends the text to: )&K as y i n Tr: M i  ,  , * If .  $p / j ^  A>  , and interprets y i n  'nightmare.'] means 'to sleep but not be satisfied' [i.e. to feel drowsy?]. It consists of  jjfc  meng ffi 'dream' abbreviated signific, and mi  phonetic.  [Duan em: M l / | | |  means 'to have nightmares during one's sleep.']  Analysis The present graph seems to contain a person element similar to that in l£o. -^T 'old.' However, in view of its componential similarity to the SW character mi ^ it seems best to regard \ $ as a variant of meng  pI^  / ,  .  Usage S gives two different transcriptions, one of which assumes that the inscription is incomplete. Due to the fragmentary nature of the bone, it is hard to know which reading is correct For the sake of argument, I shall give the complete reading:  i •• *  111  E  tfe & . (Qianbian 4.15.4)  Tested: Our household's former (i.e. dead) drowsy (?) servant is not after all cursing us. 1  I think that the heuristic application of graphic analysis enables us to make a guess at the meaning of the present graph (i.e. that it could be something to do with sleeping), but unfortunately the only inscription in which it occurs is rather hard to decipher.  ^or my 'emphatic' translation of wang " £ l »see Takashima 1988.  150  Part 4:  h\  The graph that I use to head this section is c 6 n g  \\  I AJ»  'to follow.'  However, this cannot be said to be signific in the graphs that I deal with here. I am simply using it as a convenient title for all those graphs that contain more than one ren A.  \  I  element. These graphs denote words covering a wide variety of human interaction.  Sometimes it simply represents a crowd of people, but ideas such as cooperation, confrontation, competition (in fact, many words that begin with the Latin word for 'with,' and this is not coincidental) are also represented by using more than one 'person' element. The following selection of graphs attempts to show how this is done.  151  JGWB 1025: cing AA  81 ^  ;1028:bJ Jrfej  S.19.3x617  JWJ 8.2687: c6ng kk  SW8a.l6a:  , 8.2693: bi  W , *8  fctj  JTjfe-  . AC ^ A  .ft...  SWDZ 8a.43a: This is the primary form of c6ng Tr: Cong kk  'to follow' means 'to listen to one another.' It consists of two people. 1  The SW definition as listen to one another' is perhaps intended to explain why this character consists of two r i n  elements, as Professor Takashima has suggested to  me. SW reserves the more concrete meaning suixing AMaggregate character (SW 8a. 16a).  l^t,  'follow' for the  Analysis The graph shows one person following another. There is some argument as to whether the graph should be identified as bj fct-» , in the sense of 'to ally oneself with.' The argument given for this is that the oracle bones frequently talk about the Shang king 'following' certain military figures in battie, and some people feel unhappy with the idea that the king should be following others when he should be leading them or at least sending them to fight on his behalf. The idea of 'alliance' is a sort of compromise between these passive and active stances. However, this is all guesswork concerning a social situation which we know very little about. JGWB hedges its bets by saying that cong kk 2  and bi  4 am grateful to Professor Takashima for informing me of the kambun kundoku reading of this phrase, aikiku nari. which has enabled me to arrive at this translation. Takashima (1987:63-69) has a thorough discussion of cong kk . He notes that some Japanese scholars interpret it in a causative sense 'cause to follow,' but shows on grammatical grounds that it must mean 2  'follow,' and only means 'make follow' when preceded by the causative verbs ling, ty  ) 'to make,' or hu,  %  (  p  ^  'order,' z&h.  ) 'to call.' He concludes (p.69) that "it was not always  152  ttj  w  e  r  e  anciently the same graph, while L i Xiaoding attempts to distinguish the forms  with the bent arm,  and ff  , as bj fcb , but is forced to acknowledge that even  with these forms only the reading 'follow' makes sense. I think that on graphic grounds there is little doubt that the present graph is c6ng M . • In the seal form, bj \, t j consists of two rgn  elements and differs from c6ng  Ik  also  only in facing right rather  than left. However, right~left orientation is not usually significant in the bone script, and I suspect that bj b b originally consists of two bj  t_  'spoons' in a row as it does in the  modern script, the spoon thus also being phonetic, though it is admittedly unusual for a graphic element when reduplicated to keep the same reading, as Professor Takashima has pointed out to me with examples such as ma.  > lfn  , ren A _  > cong  , etc.  Usage Probably in its primary meaning 'to follow,' usually of the Shang king following various military figures who were perhaps the leaders of allied tribes, against a common enemy. It is also used as an adjective in the expression c6ngyu  kk  1^  'subsequent  rain,' and as a preposition meaning 'by way of (a certain place).'  84 %  JGWB 1027: bjng  S.23.1xl2  JWJ 8.2691: bjng $  (this reference not cited in S)  the case that the commander-in-chief, namely, the king followed his subordinate commanders. At times he followed them, at others he made them follow him."  153  SW8a.l6a: \i  ,  M  n- % # .  SWDZ 8a.43a: [Inserts gan =f Tr: Bing  3ft  &  . JJL J~k  ,  ft  f£  .  phonetic. One source says: c6ng Ik -ff-  'poles'] makes bing  ©  :  M  after i r ^ - .]  means 'to follow one another.' It consists of cong  signific and jian  -  Ik  'to follow'  holding two [gan  -f  .  The character -f-f-  is supposed to be read jian. so the SW statement that it is 1  phonetic in bing -tj- is very curious. Note however that SW 14a. 12b defines jian as ping ^p"  level, and this could be intended to be a paronomastic gloss. 1  Analysis The graph shows two people connected by one or two cross-strokes across the legs, and thus illustrates the meaning 'combine,' which is the meaning it has in OBI. Ironically, the seal form breaks the bond between them. The modern form  $~  has  brought them back together again, but the original pictorial intent is no longer discernible. Needless to say, both the SW analyses are wrong, being based on the corrupt seal form. It is interesting to compare the present graph with bing Bing  ii.  (II.l.i.216).  shows two people standing side by side, thus representing the concept 'side  by side, together' (GSR 840). In soft texts, it is nearly always followed by a verb, and thus acts as an adverb, e.g. in such phrases as bingqu bingzud  'sit together,' and bingsh6u jL  ^  'ride together,' 'receive together (i.e. two  people receiving something at the same time),' occurring in the Shijing. Bing 2  ^  on  the other hand is basically a verb 'to combine,' and all its other uses can be derived from this. Since bing  j^-  can also be used as an adverb 'in combination, all together,' it  sometimes superficially resembles bing  , but some nuance of difference can often be  rhe Guangyun gives the reading "vt Jf^ ~t/j ken, and says it is also read like qian ^ Unfortunately the Jingdian shiwen does not have any glosses on these examples.  lr  2  ken.  154  detected, e.g. in the Shijing line  g  ^  > 4  ^  £  "The innocent ones among the people are (all together:) indiscriminately made serfs" (HY 43/192/3). Although the people are the subject on the surface, underlyingly they are the object of the action. Further, bing side.' As an active verb, bing  $  here implies 'mixing up' rather than 'side by may take an object, as in the bone sentences cited  below, but I have seen no examples of bing j$L  in soft texts followed by an object. If 1  there are any, they must be vary rare. The difference in usage between bing r|L bing  and  can perhaps be accounted for in terms of the introvert/extrovert opposition that  Takashima (1987:sec. 3.3, 4.1) has worked out for pre-classical Chinese. Karlgren (GSR 840) says they are cognate. Guangyun reads bing $  Professor Pulleyblank has pointed out to me that the  as pj iaj n,and pj iaj n and bing j[L h  is also a character bing  as bej rj', and that there  read pjiajrj', pjiajn ,and bejrj', of which the last reading is h  actually the same as bing JJL  . It would seem from this that there was some confusion as  to which word should be written with which character, and we cannot be sure that early texts have come down to us in the original characters with which they were written, so extreme caution is needed in unravelling the original uses of these characters. Professor Pulleyblank relates this word family to bi bt> 'combine, unite.'  U s a g e Mainly a place name, but also occurs in its primary meaning 'to combine,' with reference to horses:  ^  %  v^L  •  It should be (so that we) combine piebald horses, (or:) It should be the paired piebald horses.  lr  This claim is based on an examination of HY concordances.  (Jiabian 298)  155  $  # %%  > "t-  Ji. •  (Qianbian 4.475)  It should be (so that we) combine lao-horses, there will be no disaster, (or:) It should be the paired lao-horses. there will be no disaster.  1  These divinations concern the choice of horses to pull a chariot, in order to ensure an uneventful ride. Note in this connection the character ping~pian  'horses side by side  with one another' (GSR 824n).  85 \k  JGWB 1029: b i i J b  S.23.2xl40  JWJ 8.2699: M i  SW8a.l6b:  J b  1ft  , ijfi J&J •  XI  SWDZ 8a.44a: [Quotes classics to suggest that M i  ~ i b  A. is the primary form of bei  ffi  'back.'] TnBii  'north' means guai ^jjt- 'contrary to.' It consists of two people with their  backs to each other.  Analysis The words bei 'north' (<*pak), hei 'back' (<*paks) and M i 'to turn ones back on' (<*baks) are all etymologically related, so the problem is to decide which member of this set the bone graph depicts. The bone graph clearly does not depict 'north,' nor does it depict a person's back. Rather, it depicts two people standing back to back. It thus U am grateful to Professor Takashima for suggesting these possibilities of interpretation.  156  represents the verbal usage 'turn ones back on.' As for which word is primary, the morphology shows that it must be M i 'north,' as this is the form with no affixes. I would guess that its original meaning was simply 'back' as a locative adverb. The English word 'back' has a similar range of uses (i.e. person's back and back as an adverb, though not for the meaning 'north'), so this forms an interesting cross-linguistic parallel.  Usage Always used in the meaning 'north.'  92 J)f  JGWB 4346 (unidentified)  S.24.4x7 (excluding the name (ft ^  JWJ 8.2677: hul  SW8a.l5a:  YrL  Tr:Hul A-  fl|  $  )  .  M  j£j  ,  | ^  ft  ^  ,  U  means 'teaching is carried out.'J It consists of  JL  hjia  ;  £  fa  1%r  •  'to change' and r£n  'person;' hua V-> is also phonetic.  Analysis The graph shows two persons, one the right way up, and the other upside down. The original meaning may have been something like 'to turn' (i.e. zhuanhua and 'to change' would be an extension of this. Clearly, in order to convey this idea, both 2  ^This translation is based on the kambun kundoku reading, which Professor Takashima pointed out to me, and which is in turn based on the SWDZ commentary. Professor Pulleyblank has informed me that hul 4 may belong in a large family of words with initial *x - that have to do with 'turning.' This would corroborate my interpretation.  2  w  157  person elements have to be present. How is it then that SW has the 'upside down person' element hya ^  as a separate character? There are no textual examples of this character,  and Karlgren does not give it in GSR, starting his phonetic series no. 19 with hua 4 L> • Xu Shen tries to make out that they are separate words by defining hua and hua //(£j as 'educate.' However, not only is hua 4normally  b  as 'change' used to mean  'change,' it is also clear that the meaning 'educate' is simply an extended usage, i.e. to transform through education. I would suggest that X u Shen extracted the element hua YJ  from the characters in which he thought it occurred, and that it never existed  independendy. The character for 'change' has always contained two person elements, and has to in order to illustrate the idea of one person changing his position relative to another.  Usage Unfortunately the present graph is always used as a proper noun, especially in the name of the military figure Cha Zhi Hua £|a  £  Jit,  . Thus there are no examples  of it being used in its primary meaning which could corroborate the meaning 'turn' that I posit for it.  93  JGWB 282: jing.  5ff  ^  S.25.1xl5  JWJ 3.757: jing  ffa  SW3a.l8a: f f  ,  ? I  %%  .  — ©  :  Ai  .  . M  tt  , Xk  158  SWDZ 3a.32b: Ojlng  is a paronomastic gloss. Qiang yu  ^  means 'to  dispute.' Tr: Jing # £ j 'to compete' means 'to talk forcibly.' One source defines it as 'to pursue.' It consists of jing  'to argue' and two people.  Xu Shen defines jing -jp^  as 'dispute' in order to account for the double y £ n  'word' element (which is probably another of his inventions, since there are no textual examples of this element), but the bone form shows that this is a corruption of something else. The basic meaning of jing z/fu^  simply 'to compete.'  Analysis The bone graph consists of independently in this form. ^  doubled. This graph has not been found  (Yibian 8786, given at JGWB 5166 as an unidentified  graph) could be an example, but the context is disappointingly fragmentary. JGWB 283 identifies  7\  (Jiabian 916) as jing  be recognized as jing jing %fj  giajn  %j .  %j , and I think that the present element should also  This is supported by the phonetic similarity between  < *grans and jing  %j kiaJQ  < *krans.  The graphic and phonetic  similarity is so close that I do not think there can be any doubt that jing " j ^ is phonetic in jing %%j • But what is the primary meaning of jing for branding the foreheads of criminals' on top of rin the primary form of qing  jgj ? It consists of xin T  'tool  'person.' I suggest that it is  giaj rj < *gran 'to black-brand' (GSR 755g). The phonetic  fit is perfect. The present graph is thus a phonetic compound consisting of jing However, the two r i n A  doubled.  elements also suggest the idea of one person racing after  another. Thus although it is basically a phonetic compound, it is also highly suggestive of the meaning that it represents. I believe that this was a conscious piece of design on the part of the creators of this script. From a purely graphic point of view, the graph could of course simply be seen as two people walking peacefully together. However, the meaning  159  of the word that the graph represents suggests that the intention was otherwise. One cannot analyse the bone graphs in a vacuum, without reference to the meanings of the words they represent. JWJ includes the graph J J unidentified.  (S.40.1) as a variant, though JGWB 3005 lists it as  A similar variant is also found in bronzes (see O/NJWB 304/389).  However, the form with two rin  A. elements is the commonest form, and it is from  this form that the modern character develops. It is not clear to me why there should have been a variant with the djl  ^  element.  U s a g e Personal name. Also a verb in a sacrificial context, meaning unclear.  94  %  S.25.1x8 95 ^  JGWB 1033 (both forms): zh6ng  $i  S.25.1x65 (excluding certain contexts)  JWJ 8.2703 (including ^  SW8a.l7a:W  ,  Y. zhong ffi  A  .  &  Tr: Zh6ng )$K 'crowd' means djio. ^ crowd' and nui @  , Q  , )$L ^  .  'many.' It consists of yin  'to stand as a  'eye,' thus expressing the idea of 'numerous' [presumably because  there are numerous eyes in a crowd].  Analysis  ^  160  The graph shows two or three people standing under the sun, thus illustrating the idea of 'the masses.' The sun suggests that they are out working in the heat of the day, and indeed the OBI context suggests that the body of people designated by this term was primarily a workforce (see Keightley 1969). In the seal form, the sun has been distorted to 'eye,' and this is also the case in the bronze form (see O/NJWB 1117/1376). This further supports the idea that they were a workforce, with the eye representing the supervision (literally!) that they would be under. L i Xiaoding recognizes the variant ^\\\ , with no sun on top, in the inscriptions Tieyun 231A and Qianbian  5.20.2.  S transcribes both with the sun on top. A n  examination of the rubbings shows that in both cases the bone is broken away immediately above the \ \ \  graph, so there is no way of knowing whether there originally was a sun  element on top of these two graphs or not. JGWB 1032 records one instance of ty\ (Jiabian 2858). identified there as the SW character yin of the rubbing reveals that the rightmost rin  ^  . However, an examination  element is somewhat separated from the  rest of the graph, and is most likely quite unconnected. Thus what we really have here is c6ng kk , and the context supports this:  ^  -4  s  S Xk  ...  (Jiabian2S58)  If we dance, today there will be subsequent [rain]. 1  lr  The author of JGWB (Sun Haibo) himself says that c6ng  kk  here is used for z6ng  as in z6ng  yu 'loosed rain > pouring rain.' The basic meaning of zdng is 'to let loose' (GSR 1191h), and the idea is that the Shang expressed the idea of torrential rain by saying 'loosed rain.' The fact that I have been unable to find such a usage in soft texts does not totally preclude the possibility that the Shang had such an expression, but on the other hand taking c6ng kk The expression cdng yu  kk  ffi  at face value also makes sense.  nearly always occurs embedded in a nominalizing formula:  ^ kk 1^ (positive) or ~ t kk ^ (negative). It also usually occurs in the apodosis of a two clause sentence, where the protasis specifies some ritual activity (e.g. dancing), the aim of which is to obtain rain. 'Following rain' would then be the rain that follows from the ritual act.  161  This phrase is repeated several times on the bone, and in all other instances we have the graph cong ^\ I /A  . I think it is safe to say then that the character zh6ng ^  has  always had a sun element on top, and it is an integral part of its meaning: the graph does not simply depict a crowd of people, but a workforce, and their role in Shang society as seen in the oracle bones fully corroborates this.  Usage As already mentioned, it refers to the zh6ng-workforce.  96  TH  JGWB 4532 (unidentified)  S.25.3x2  Not in JWJ  (fe , )f. % .  SW 9b.7a:  t  SWDZ 9b. 17a: SW says Tr: Shu  ft  J*L>  1%^ "~f  .  AA  in order to explain the  means 'a crowd under a roof.  1  is the old form of the character guang also represents the idea of zhongsheng  &  J*~ on top.  It consists of yan  'roof and  ;  'light.' [Xu Xuan explains that 'light' 'numerous and flourishing.']  Analysis The graph consists of ^ \ shix Jim  , identified by Y u Xingwu and Chen Shihui (1959) as  , on top of three people. It is probably the primary form of shu  in the  162  sense of 'numerous,' especially with regard to people, as in shur6n  yC  commoners.  dziajk  Qia < *+aks (?) itself contains shi  / ^  h  < *dak 'stone' phonetic, as Karlgren suggested (GSR 804a, but he did not feel confident enough to include it under sM  phonetic, GSR 795). Y u and Chen say that shu Jjfc  is the original character for zhu "^r tpfa' < *ta' 'to cook by boiling.' This claim is based on the office of Zhushi commentary says that shu  t^/  in the Zhouli (HY 10/7a). The Zheng Xuan  here should be read like zhu  as in y a o z h u  'medical decoction.' Biot (1851.11:386) transcribes accordingly Tchou-chi (i.e. Zhushi), and translates the title as "Cuiseur (d'herbes)." The role of this official was to decoct pesticides, so this supports the idea that the primary meaning of the character shu  was 'to boil.' There is also a character  where it says that the people of Wu  ^  in the Yanshi jiaxun (ap. Kangxi),  use it for zhi ^  tciajk < *tak'to roast.'  The three person elements have the same function as in the character zhbng J^k  }  i.e. to indicate a body of people.  Usage Possibly in the sense of 'numerous':  (Qianbian 5.25.1) Gengxu-day cracking, tested: There are many locusts, it is that God is 1  ordering dearth.  lr  The graph ^ *  has not been identified with certainty as an independent character, but it is generally  agreed that the expression element in $f|l  2  (S.247.1) means 'this autumn,' and that  Jj|  / -^I|L> , which SW (7a.l8b) gives as the zhduwen form of gju  JWJ 7.2369 and 13.3939). SW (10a.20b) says that the phonetic element in aiy.  ^m.  is the phonetic 'autumn' (see  is jiao  163  97  1$\  Not in JGWB  S.25.3xl  Not in JWJ  SW 8a.l7a: $  , <fr  &j  SWDZ 8a.45b: [Inserts — Tr: I i ^ ^ b  . &  &  ,  before  fa  .  %  ^  .]  means 'to gather.' It consists of yjn  'stand in a crowd' signific and qu  phonetic. [One source says:] A village is called a 'gathering.'  Analysis The bone graph may be analysed in the same way that X u Shen analyses the seal form, i.e. ai  3pL  phonetic and three people indicating a crowd. Qji Jfe-  ts'ua* <  •ts'aw? means 'to take,' and it seems likely that it is not simply phonetic in the present graph, but that it is in fact etymologically related to ji ^  ^  'burn a turtle shell but fail to crack it' ( "f^d compound of gui  'turtle' and huo  ^H  dzua' < *dzaw'.  ). which it analyses as a semantic  fire.' However, the bone graph ^  stands for gin.  without any 'fire' element added, and is both graphically and contextually distinct from gni / j3P  'turtle.' In most cases the graph J»  is drawn with what appear to be wings on the  back, and feelers on the head. An example from Cuibian 4 is particularly insect-like: y yj . I t probably depicts an insect that is harmful to crops, e.g. a locust. It was probably originally phonetic in jiao , but later was corrupted to the graphically similar gni Understanding as gi£a 'lack,' though this interpretation is uncertain. At any rate, it probably refers to some kind of calamity.  164  Asinzh^ng ^  andsM* ^  , the three people represent a crowd.  Usage The single occurrence supports the meaning 'gather':  * f i t • £ *%. ^ t f  4-  . i *  . (Qianbian 4.27.2)  Should not si-sacrifice. Perhaps it should be the Lesser Officials that we assemble and order. The king will not regret it.  98  JGWB  349:  di>u H  S.25.3xl6  J W J 3.889: d5M |*J 5  SW 3 b . 7 a : f f  SWDZ  ,  #L  3b. 15a: [Emends text to:  Tr: DQU. ( | JJ  [SWDZ  |  . *  , S ?i ^8 f i  «  i  ,| 7^.] t  consists of two men opposed with weapons behind. It depicts fighting.  em: D&u.  means 'to struggle.' It consists of two j i  each other, and is a pictograph.]  Analysis  ,  $L  elements facing  165  The bone graph shows two people having a punch-up. It is pure fisticuffs, no weapons involved. X u Shen saw weapons in the seal form because he regards it as consisting of j i  ^(j  doubled, an element that he thought depicted a hand holding  something, though here again he is mistaken, as the bone form kneeling with empty hands held up. As you can see, d6u ^  shows a person does not come from ji  : this is merely the way in which the seal script has standardized it. Ki"  Note also the tousled hair, which Ye Yusen describes as nufa standing on end with rage. The same sort of hair is found in ru5 ^  /  1  lao  /  'old,' and fli ^  / ^  'wife.'  'hair 'to agree,'  It is hard to find a consistent  symbolism behind it. Professor Takashima has suggested to me that the grapheme  ^  may have had a 'double function,' the graphic equivalent of the sort of antiphrasis found in such pairs as sh6u  dzuw' 'receive' and shou 4 &  may have signified wild hair in d6u l^f  dzuw 'give.' h  but smooth hair in ruo ^  Similarly, $  .  Usage Place name.  1280  *ir  JGWB 4321 (unidentified)  S.194.2xl  JWJ 7.2373: & l t (not in SW)  SW7a.l5b:JjJL*g'.4*F %  £ |g-MlL • ^ £ ,£ f •  S W D Z 7a.43b: In the Huainanzi it is written H ^ wntten  or 4 & . All refer to wild rice.  . Other books have  , also  166  Tr: Li HfL means dao^E? 'rice.' It refers to the perennial variety. The character consists of h i s?~ 'cereal' signific and ni  phonetic.  Analysis L i Xiaoding transcribes the right element as if it were bji  , but actually it is  slightly different. Although the two figures are back to back, the left figure is higher than the right figure, as if riding on its back. The graph * jf^ has not been found independendy in OBI, so I shall identify it as n| It  and discuss it here.  S W 8 a . 2 7 b ( f c , lit Hit*-.  /JLJ*  :  , t # .  Tr: Ni IL means 'to approach from behind.' It consists of shi F  1  hi  ['body'] signific and  *L. phonetic.  Analysis The SW definition 'approach from behind' seems to be just what the bone graph depicts. Note also the graph 1  , which unfortunately only occurs as a person's name  4 (S.324.1), but which is probably the present graph augmented by the chi  |  'walking'  radical, thus emphasizing the meaning 'to approach.' (JWJ 2.535 identifies it as chi x£. 'slow,' but I think this is mistaken, since chi *~L phonetic, which consists of wei ^  contains x i  'tail' on top of n M  'ox':  'rhinoceros' the present bone  graph does not contain such elements.) Various phonetic compounds of nf jfc. have the meaning 'close, familiar,' and these meanings are clearly related to the meaning 'approach.' The graph shows one person who is so close to another that he is almost on top of him. The corruption of the left-hand person into shi f* and the right-hand person into bj was probably structurally conditioned.  167  2885 Iff  JGWB 827: A  S.444.3xl5  JWJ 7.2227: li  SW7a.8a: f$(  , %  3- ~$ JL  %  .  JX jL  , JI  Tr: Lji -VK ; a group of five hundred men in the army is called ly. 'fluttering streamers' and c6ng XL  van  ; cong JA  corruption of cong XL  ;  JX ,  sjfc . It consists of  represents 'togetherness.'  is the old form of hi ^ K . • In ancient texts it is used for Lu  The top part of the SW old form is zhi jL  XX  as in L u and Wei  , while the lower part is probably a  , as Luo Zhenyu (JWJ 7.2227) surmises.  Analysis The graph shows two people standing under a flag, two standing for many. 'Troops' may be the original meaning. As in zhong  , the two people show that the graph refers to a group of people .  Usage In its original meaning 'troops.'  ^he reason Xu Shen phrases himself thus is to make it clear that the old form of lu ^ tJ£  was only used  for hi ^ as the name of the state, not as an ordinary word. As an ordinary word, it means 'dull, blunt, simple' (GSR 70a).  168  Part 5:  In the bone script, as a graph component, it is very hard to distinguish this element from ren j  . As Karlgren says (GSR 561a): "The archaic graph is practically identical  with that of  ." There are one or two graphs that possibly contain it, but it is very hard  to be certain, and it has therefore seemed to me rather poindess to discuss them. It is however important to aver the fact that such an element did exist, since it is the origin of the shi  radical, whose general significance is clearly that of 'sitting' or 'being located,' as (now written %. ) 'to sit.' It denotes sitting as on a chair, and thus differs from  the element J i  , which shows what I assume was the usual Shang way of sitting, which is  actually kneeling. The rarity of the /  element compared to the i i element in the Shang  script shows that the latter was their usual way of sitting. Notice that the phonetic element is always written underneath the shi f* element, thus implying that this is what the person is sitting on. This is typical of the way in which the designers of the Chinese script designed their graphs, as I have already had occasion to mention. In some cases, r£n ^  may have evolved into shi f*  'to defaecate' and niao f[  variant, e.g. in the case of shi Thus the occurence of the shi P  simply as a structural I  'to urinate.'  radical in a modern character should not necessarily be  taken as an indication that the action denoted by it was originally performed in a sitting position.  In the example of p i ^  , the s_hi j  2  element is actually evolved from the  169  kneeling figure ^  . Its evolution into shi j  3  was thus purely a structural adjustment to  accommodate the kou 12 element. The use of shi / * for y i ^  in some early texts is not, as some have supposed,  because the barbarians sat in this fashion, thus distinguishing them from the Chinese, who knelt on mats, but is rather a phonetic loan (shi T  5i < *+aj, y i ^  j i < *1 b]).  I would  imagine that in Shang times the surrounding tribes sat in the same way as the Chinese, i.e. in the posture illustrated by the graph \  .  Although hard to find in the bone script, the shi  element is well represented in  the bronze script (see O/NJWB 1146-1150/1410-1421-though most of these are not in SW).  Notice that we have shi )  1150/1421), but in the OBI form ^ of r6n \  evolving into shi f*  distorted the ren \  into shi j*  in w j i  here (in the character QU_ $  we simply have rin ^  , O/NJWB  , so this is another example  as a structural variant: having to make room for the tail .  In this part then I shall deal only with the graph that has been identified as shi T  ;  170  3  j  JGWB 1048: shi  j  2  S.5.2x21  JWJ 8.2745: shi  SW8a.26b: f  f  ,  f|  | < i?r  JL&j  . ft ...  means 'to lay out.' It depicts lying down.  Analysis The graph shows a person sitting, as if on a chair. The primary meaning must have been 'to sit' in this fashion, though it is very hard to find remnants of this usage in texts. Wu Dacheng (ap. NJWB 665, not quoted at OJWB 539) says that in the classics, •> Shi/ "^  2  means zhu 3~ 'to preside,' and in the opening phrase of %. T  ," Shi wM T 46  means ji yvM |?p  Shu.Wuzizhige  4£ 'succeed to the throne.'  Unfortunately, this chapter is found only in the jinwen text, so I am unable to offer Karlgren's opinion on how it should be translated (he only translated the guwen text), but I think it would be reasonable to render it as When Tai Kang sat on the throne.' Karlgren translates the Shijing line  jl <yCjj[£  f"  (HY 66/254/5) as "The  good men sit motionless and silent," and derives this from the meaning 'act the corpse' (i.e. play the role of the representative of the deceased at an ancestral sacrifice), but perhaps here also we could simply regard shi 'the good men sit.'  f  as being used in its primary meaning 'sit,' thus:  In ritual texts, such as the Yili, shi  usually refers to the  representative of the dead. The SW definition 'to lay out' is a loan usage.  In order to account for this  meaning, X u Shen suggests that it depicts a person lying down. Since the rin \  element  sometimes represents a person lying down, one might be tempted to consider the same possibility for shi  f  . However, there is no evidence that shi  f*  ever meant 'lie  171  down.' There is only evidence that it meant 'sit.' The meaning lay out' should rather be understood as 'set forth.' The relationship between the two uses is exactly the same as between English 'sit' (intransitive) and 'set' (transitive).  Usage It occurs as the name of an enemy tribe, so it would seem to be used as a loan for yi 4 e > 'barbarian.'  172  Part 6 :  The element * > J  *  \  has not been found independently in OBI, so I mark it with an  asterisk to show that it is a hypothesized form. Apart from the graph which S.28.1 lists as a variant of the unidentified grapheme W graph, i.e.  (Kikko 2.21.18),  , it occurs in only one  . However, Tang Lan, in his exposition of \ \ (JWJ 14.4143), proposes  to identify * \  with the graph \f  that SW (4b.5b) gives as the old form oftijjrn^ , and  suggests that it is the primary form of dian i l l , which Karlgren (GSR 375m) defines as: '...(fall on the head:) fall down, be overthrown, overthrow.' Tang hypothesizes that the same element is phonetic in zhen  . This would make the SW (8a. 15a) claim that zhen  IJ| contains 'person upside down' correct, except that it would not be hua ^  signific  but the primary form of dian Ijll as phonetic. A phonetic problem that has to be discussed here is: which O C rhyme does the zhen 4  phonetic series belong to? In order to support Tang's claim, it should belong to  the zhen ! j | rhyme.In favour of this classification is the fact that SW (9a. 8a) gives JI. JL  fa:  as a variant of zhen 4 , and this variant, as you can see, contains zhen  phonetic.  Although he does not mention it explicitly, this is perhaps the reason why Wang L i (1937:67) puts the zhen  series in his jian jfc. type (i.e. zhen ^  rhyme), and Tang  Zuofan (1982) follows this. However, Tang Zuofan lists the character tian  as w6n  yL rhyme, and in a later study Wang L i (1937a: 133) classifies the zhen  series as  173  zhun 5,? rhyme (i.e. w6n ^_ rhyme), giving as evidence the fact that zhen with yjin  in the Shijing (HY 77/290), and zhen  The only other rhyming occurrence of a zhen  rhymes  with rgn § > in ChuciXisong. (  series word in the Shijing is tian  with mei >fci (HY 9/43/2), which Wang L i (1937:67) classifies as jin 7f  ^  type (i.e. w i n  >C rhyme) and Tang Zuofan as yuan 7T_> rhyme. Karlgren (GSR 453) classifies the zhen^  series as w6n  rhyme, and indeed this is what the rhyming evidence points  to. Thus the evidence provided by variant characters conflicts with the rhyming evidence. In spite of the phonetic problems, the fact that SW gives ^  as a variant of zhen  is significant, and I am basically in favour of Tang Lan's identification of the present  element as the primary form of dian  .  174  1124  | ^  JGWB 1643: f M $1  [=zhui $  'fall']  S.178.4x2  JWJ 14.4143: hui f §  SW 14b.2b:  (GSR 11a reads djm)  f$,%t%&&.  , £  SWDZ 14b.4b: Commonly written ^ TnDui  andsjii ^  and pronounced zhui.  f  \  'wall with steps' [this is 1  , but this is clearly what the bone graph depicts] signific  phonetic.  SW 14b 3 a j | ,  t & rt. IX i  9k  it * * i. % , & - £  and  .  means 'to fall from a height.' It consists of f i  not the SW definition of f i  Tr: Hui  f  ft  , §. # . [ g * -fc 2L ,  analysed as two ZU&  % 0  :  ~ A.  : to destroy a city wall is called hui j " ^ . It consists of f i phonetic. [Xu Xuan et al. note: There is no character ^  &  \  'wall' signific  in SW, but it may be  . The strength of the multitude assists therein (the destruction of  the city wall), therefore it contains zuo ^£ 'to assist' doubled. It is now commonly written  , but this is incorrect] The seal form is ifrf [i.e. j f g ]. According to SW (SWDZ edition 4b.30b), hui fai is the abbreviated phonetic in  sui fj|  . This suggests reconstructions for these two words something like * x a l and w  *7r al respectively. But why did the initial in the former fail to palatalize and front to s- in w  E M C ? Note that hui  ( |}f"  )*  s m  ^  a c ta  grade IV word in a chongniu rhyme, E M C  xjwia, and it is possible that in this case the E M C medial -j- goes back to an O C * - l - , so we should actually reconstruct * x l a l . w  2  This still leaves the problem of how zuo  H am indebted to Professor Takashima for suggesting this translation. For the suggestion that it may have been a medial *-l- that prevented palatalization of the initial in certain grade IV words, see Pulleyblank 1984:176. 2  175  *tsal? can be phonetic in hui $ | reading dud, as if from *lwal  * x l a l . Karlgren (GSR 11a) gives hui ?|L w  the modern  but gives hui i^SL both the readings du6 and hui (GSR  lie). It seems to me that the reading du6 should be reserved for the intransitive verb 'to  it fall,' since this is the reading of du6  'to fall,' and the reading hjii should be assigned  to the transitive verb 'to fell.' The phonetic relationships are highly problematical. Professor Pulleyblank has informed me that he thinks it is possible to connect the initials in zuo f_  and sjn  , though he has not yet worked out the details. He compares zong  tsown, containing xiojuj &  xuawrj phonetic.  The initials in duo and hui are  ?! admittedly far apart, but the fact that the character ^  has both these readings suggests  that they can be reconciled in some way (though perhaps the answer lies in prefixes). Analysis The graph shows a person upside down next to a wall, implying that he is falling off it. JGWB and JWJ also both include the graph £ ^ (S. 179.1x4) as a variant, which shows a child falling. Perhaps a child is more likely to fall off a wall than an adult. It is possible that they are indeed variants of the same graph, but the inscriptions with ttl in are too fragmentary to determine its usage. The present graph, on the other hand, occurs in a nice long inscription (Jinghua 3) that tells us how Prince Yang fell out of the king's chariot during a hunting accident. Thus the meaning is clearly 'fall.' L i Xiaoding (JWJ 14.4145) suggests that the upside down person a phonetization), and this zuo  was corrupted to zuo t  was later doubled to %. .  I £  (i.e.  This is an interesting  hypothesis because phonetizations, naturally, are often not such good phonetic guides as original phonetics, so this would explain the divergent initials of zuo ~& and duo~hui . If this is true, it is a great help, because it means that we do not have to account for the divergence in the initials. One can further suggest that $ui • $ fji  abbreviated, but simply *  doubling in hui  *JEE.  does not contain hui  , the phonetized form of -j ^ . The reason for the later  is not clear, but there are other examples of this sort of doubling in the  176  bronze script (cf. geng which the bing  , O/NJWB 415/518 and & } § : , O/NJWB 1812/2319, in  and hu T \ elements are usually doubled). The reasons are perhaps  calligraphic rather than linguistic. That is to say, to make the graph look more pleasing to the eye (in the subjective opinion of the scribe), rather than to reflect some phonetic difference between the simple element and the element doubled. L i Xiaoding also suggests that du6 f ^  (which X u Xuan regards as the proper  ?I form of duo *jr 'to fall'-see SW 14b.3a) is an alternate graphic development from the same bone graph, thus: \]^ ^  >f^  > f^  • However, while the corruption of  ^  >  is understandable in terms of phonetization (which L i does not mention), a corruption  to xi $  is unmotivated. It would be more likely that the form hui  to dub f \  changed direcdy  -this would be motivated by (1) getting rid of the non-character ^  , and (2)  A getting a more accurate phonetic, since the initial of duo & would be closer than that of zuo  . Thus Li's proposal that there was a development f |  > f&  It explains why there is no character * %i , and why zuo ^  > f^r explains a lot.  and hui  have such  divergent initials. His proposal is further supported by the phonetization process that one witnesses so frequently in the evolution of Chinese characters. The identification of the present graph as hui 1^ dui  may thus be accepted in preference to the JGWB identification as  . However, whether the graph ^ |  is also to be identified as hyi ] ^  be seen. L i says that this form became extinct, though "£ developed into dian * ^  ZJIQ  ^  rather than as a phonetic compound with * y  of •)  and ^  ,  as phonetic. As for his proposal that * y  , the bone forms (JGWB 867, also 3456, 3868 and 4742) clearly  do not have 'person upside down' but M t  Usage  could also perhaps have  . Tang Lan proposes to identify the present graph as a variant of  /lj|l| . but I think it is best to regard it as a huiyi ^  is phonetic in zhen ^  remains to  'spoon.'  177  In its primary meaning 'fall,' as in the well known inscription:  %  t T,  tl  I : >S) -t  . i  if) ®  (Jinghua Divining on guisi-day by cracking bones, (the diviner) Que tested: There shall be no misfortunes in the (next) ten-day week. The king, having interpreted the omens, said, "There will thenceforth also be harm in this (omen), and (will prove) to be as prognosticated (by me)." When on jiawuday the king went chasing wild buffaloes, a minion took charge of (steering) the chariot horses; (he) drove the king's chariot precipitously, and Zi Yang [who was riding the chariot] also fell down.  here is presumably read du6 rather than hui.  rhis translation is taken from Takashima 1984:32, inscription B3 (a).  lr  1  178  1  Part 7:  In the present group of graphs the emphasis is on the hair, and the r i n  \  element tends to be neutral in its significance, as one might expect, since it is not the focus of these graphs, but rather a carrier for the hair which is the focus. Long hair is used chiefly to suggest 'age' or 'growing,' as in lap.  / "7t 'old' and zhang A  A  % >k 'elder; to grow.' It is also used for other purposes, e.g. in wei fineness of hair, in pi ^  1  I T u it suggests the  / ^jr it represents the long hair of a grown woman, and in xu  'beard' it represents hair itself with no further symbolism. In ku_ 'to wail' and dou ^c^f / T€.  v  through distress, in dim  I  'to fight,' it could perhaps represent tousled hair (in ku l^)  through fighting).  179  33 S.ll.lx8 41 S. 11.4x8 672  &  JGWB 1046 (aU three): Mo.  S.108.3x8  JWJ 8.2739 (all three): M Q  SW 8a.25a: &  ft  ,  ^  6  SWDZ8a.67a: ^  •£  A . . -t +  8  .  M  A  , 4  ,  £  .t  *I  . /C ... is not rin  A  , but part of m£o  . [The bone graph shows that  he is mistaken.] Tr: Lao consists of ren  means kao -fe  'old.'  "When he is seventy we say, 'He is old.'"  'person,' mao  'hair,' and hua  ^-  1  It  'to change.' It means to  say that the hair turns white.  Analysis This graph, as Ye Yusen so eloquently describes it, depicts a bent old person with long hair hobbling along with the aid of a stick (  A  Jt^  ^  \\$*.  'fir  ). It is interesting to compare the Egyptian hieroglyph for 'old,' which also depicts a - - — i  man bent over a stick: > J^j ((Gardiner 1957:444, sign A19). One should also note Serruys' suggestion that the hair, m £ o , also acts as an 'endomorphic phonetic' (Serruys 1957:153). I think this is highly possible. In the bronze form of !a2  ,  ^  (O/NJWB 1138/1402), the stick has already  been corrupted into the shape that it has in the seal form, and which Xu Shen identifies as ^Liji.Quli H Y 1/8. Translation from Legge 1885.1:66 (|27).  180  his huJt  £  radical. Far commoner than 1|Q  examples) is the character kao -%}  (O/NJWB  in bronzes  (O/NJWB  1142/1406 lists 107/134 examples), where  it is used in the specialized meaning 'deceased father.' The element understood as kao  3"  lists five/seven  phonetic, which is homophonous with kao  ~>~  is usually  . One can see  that the bronze scribes intended it as phonetic, as sometimes the phonetic alone is used to represent the word (e.g. on the vessels J|[  j £ - ^]  ^  and  4ty  ). However, some forms also look very much like the bone form of lag.  -4$ , with  the old person leaning on a stick, e.g.  I think it is quite possible that the bronze form of kag. bone form of lag  is the direct descendant of the  , and that the stick was corrupted to kao  "5"  as a phonetization.  Before this phonetization, the same graph was probably used for both kag and lao, and the two words are of course etymologically related. S W uses them to gloss each other (though in practice there was a difference, lag.  Jk.  being used for 'old' and kao  'deceased father.')  Usage Place name.  34  ^ ^  S . 11.2x34  J G W B  4300 (unidentified)  J G W B  4378 (unidentified)  for  181  36  ^  JGWB 4299 (unidentified)  S.11.3x30  JWJ 9.2967: (34 only) chang  SW8a.7b: ^  ,  ty  A>  •  ^  A  Tr: Wei 7 ^ 'fine' means miao "^jr  Tr: Wei  . ?!-  ^  ^  £-  ,  *  f  • 'person' and pji  abbreviated phonetic.  .  , jfe $  AL *  nieans 'to act secretly.' It consists of chi  implying 'action'] and wei  £  'subtle.' It consists of rdn. .A-  X_ 'hand holding stick,' with 41  SW 2b.9b: #f  ,  ^  .  *  #  0  :  signific [a walking signific,  phonetic. The Chunqiu says: "The Duke of Bai [fled to a  hill and strangled himself] but his followers concealed it."  1  Analysis L i Xiaoding includes the present graph among his variants of chang believe that only the form Houxuan's identification as  ^ fu  is correctly so identified (see 1.7.39). (JWJ 9.2968).  has the meaning that SW ascribes to wei  I follow Hu  This element does not occur as an  independent character, but only in the aggregate forms wei definitions I have cited above. Only the form wei  , but I  and wei 4njli whose SW  is current in modern Chinese, and  . The definitions 'subtle' and 'secret' are  obviously different aspects of the same basic meaning, so I think we can be quite confident here that we are dealing with one and the same word. The basic meaning of wei  !  H Y 494/Ai 6/Fu 3. X u Shen has left out some of the text, which reads in full:  {g  ^  M y translation is based on Legge 1861.5.2:847.  is  182  'small, fine, slight, weak,' and I think the present graph represents this by emphasizing the long fine hair on the person's head. This is particularly clear in the graph  ^  (JGWB  4299) which may be regarded as a variant. L i Xiaoding overlooks this graph, though he transcribes the same element with a foot underneath existent character j |  (JWJ 2.463).  correctly identifies as w^i  ^  (JGWB 4299) as the non-  Note also the graph \ \  , which JGWB 998  .  Usage All the above mentioned variants are used as a place/personal name.  38 ? [ U  JGWB 3238 (unidentified)  M  S.11.4x3  JWJ 2.431: ky  SW2a.l6b: £  ^  ,  %  &  .H  W  % %  , %  . K  ...  SWDZ 2a.30b: [Expresses doubts over SW's 'abbreviated phonetics' and suggests that ku_ •fC  v  originally referred to the howling of dogs, and was only later applied to the wailing  of humans.] Tr:Ky  ^  v  means'the sound of wailing.' It consists of xuan 0tz signific [defined at  SW 2a. 16a as jinghu %  'a cry of alarm'] and y i  abbreviated phonetic.  Analysis As L i Xiaoding notes, the present graph occurs only in a very few inscriptions with little context, so Ye Yusen's identification as ky ^  N  must be regarded as tentative. Ye  183  describes the graph as depicting a person beating the breast and jumping (piyong  jffi  ), while the two mouths represent wailing (according to Chinese ritual texts, this was the appropriate way to express ones grief at a funeral ). If Ye's identification is correct, 1  then this means that the original graph for ky -jC contained a person rather than a dog, and the person was corrupted to quan "^C dog in the seal form. X u Shen tries to account for the dog by claiming that it is y i  abbreviated phonetic, while Duan tries to explain  it by saying that the word originally referred to the howling of dogs. If ky.  ^  originally  contained 'person,' then these explanations become unnecessary. If the corruption from a person to a dog was not purely graphic, it is hard to see what the motivation could have been. If, as Duan maintains, it originally referred to the wailing of dogs and later came to be applied to people, one would have expected a corruption the other way round, from 'dog'  to 'person.' Ye's description of the central element ^  as 'beating the breast and jumping'  does not seem to be accurate. The hand is not turned to the breast, and the foot is not even depicted. The focus is rather on the long dishevelled hair, which I think is probably intended to suggest the state of disarray that one's hair would get into while mourning, especially in the vigorous fashion recommended by the ritual texts. The two mouths would of course represent the wailing.  Usage  A good context is provided by the following inscription:  A * i  :  %b ,  % t- iC> .  (Qianbian 5.10.7) Wuwu-day, tested: If we wail and cry, there will be approval and no blame.  ^ . g . Liji.Tangong: $ [sic] H , %_ i L % A , (HY 4/15) "Beating the breast (by the women), and leaping (by the men) are extreme expressions of grief (Legge 1885.3:169 |28).  184  I have taken yi  as standing for ai  , since the context seems to make this reading  appropriate. However, the context is really too limited to be sure of one's interpretation.  39  ^  JGWB 1133:  ching -JL  S. 11.4x2  JWJ 9.2967: chang  •L  SW 9b. 13a:  . fL I'l ft /ft. . -t. ¥  &  . /?  , /ff.j .-t AL .  Tr: Chang -Jj^ means long,' both of time and space. It consists of SIR T C J 'high with the top level' and hui.  ^-  'to change.' W u ~R-> here represents the idea of high and  far. When something goes on for a long time, then changes occur [this explains the role of thehua  element]. Wang  Xuan et al note: Wang  T^L is phonetic.  (¥  is wang  upside down means buwang ^  this represents the idea 'enduring.']  7\.  upside down. [Xu T H 'unperishing,' and  is the old form of c h £ n g •k  .  \  is  another old form of chang f^, .  Analysis X u Shen was clearly non-plussed by the seal form in front of him, and invented the most fanciful explanation imaginable. The bone form is just a simple graph of an old person with a walking stick. The second 'old form' given in SW is strikingly similar to the  185  bone form, so the identification is quite certain. In the seal form, the stick, arm and leg have become corrupted into a meaningless mish-mash of lines, so it is no wonder X u Shen was unable to see what lay behind them. The graph is very similar in construction to 1£Q ^  / -^jr . However, the top part is distinctly different, and lacks a definite explanation.  Karlgren suggests tentatively that it might represent "long hair or a tall, plume-like headdress" (GSR 721a). The horizontal bar suggests that, whatever it is, it is on top of the head, rather than part of the head, so it seems more likely to be some sort of distinctive head-dress that was worn by elders in the community where this graph was created. The graph thus probably represents the meaning zhang 'elder,' perhaps evolved from an earlier meaning 'grown up' (since zhang also means 'to grow').  Usage There are only two occurrences in Shima, and one is fragmentary. The other reads:  (Houbian 1.19.6) Perhaps when offering to the Eldest Son, it should be Yuan who proffers , 1  2  (then) the king will receive divine aid.  Although the precise meaning of this inscription is unclear (my translation is highly tentative), at least the collocation zhangzi encourages me to think that the graph depicts an elder.  ift would seem from this inscription that he has died.  is in terpre table.  This further  186  43  JGWB 3602 (unidentified)  3p  S.11.4xll  JWJ 12.3599 (including ^  SW 12b.2a: f  ,1 J f *  ^  , ^  , ^  ^  ^  2  ^  3  ): ai  M  S W D Z 12b.5a: SW gives no old form under gul make che  V  4,  phonetic, but this is impossible: che  *ts sj. In fact the £h£ Tr: Oj  V*  ^  ,U  . [Duan also emends the text to W  trMat < *tr'at, gj. 4r  ts'ej <  element is simply a standardization of the wife's hairdo.]  is the lady who is equal to the husband [i.e. 'wife']. It consists of nii  'woman,' che  V  'sprout,' and y6u  4  9^  consists of  R  and ny.  >*  'hand.' The hand represents the idea of  running things [i.e. running a household]: this is the wife's job. ^ •3r  9  : the old form of gi  is the old form of the character gui  ^  'noble.'  Analysis As an independent graph, the  n  element is a variant of 1£Q_  present graph it has to be recognised as a variant of "f  , but in the  , as L i Xiaoding does, since the  context, although fragmentary, indicates the meaning 'wife':  %%  ... (Yinxu 651)  ...Ancestor Xin's wife... ^Qianbian 5.17.4 (S. 137.4x1). S.137.3xl3. S.137.4x2  2  3  187  (Waibian 46) Guiwei-day tested: Seek birth from the wife Ancestress Geng.  The graph may depict, as L i describes it, a hand tying up a lady's hair, or else perhaps inserting hair-pins. He quotes the adage  2,  ^  ^  "J  ^  ^  ^  'when a girl has reached the age for wearing hair-pins, she can become a person's wife.' It had the same significance as capping did for boys. These customs are mentioned in Zhou ritual texts, and perhaps the Shang had similar customs. The hair-pins are not actually 1  illustrated in the bone graph, nor in the bronze graph as an independent character (see NTWB 1956, not in OJWB between 1534-5), but they do appear to crop up in qi -*r the phonetic element in *  , a bronze variant of ji  as  (O/NJWB 643/787):  ^5 Note also fjfl  , which occurs on the Stone Drums (in the Tian Ju ^5  ^  ode) for ji  (Mattos 1973:279-280)  Karlgren (GSR 592) says: "The upper part in the character is...[gX] ^ not simply phonetic, for  dz'isr 'equal' and  ^ . g . Liji.Neize: \  vrp ^  K37), n  .... But this is  ts'isr 'consort' are cognate words,  "At fifteen, she assumed the hairpin" (Legge 1885.3:479  "T" v f p ^ ! "At twenty, he was capped" (Legge 1885.3:478 f34).  188  the consort being the one wife who is socially the equal (the "mate") of the husband. The element at the top...is a drawing of the hair-pins characteristic of the married woman." However, the bone graph ^  simply depicts a woman with long hair, which in this case  should not be taken as a sign of old age, but of womanhood. Although the hand element has been interpreted as doing the hair up, one should note that a hand over a person in OBI usually indicates the subjection of someone inferior, I ^  as in M  ( $1SL ) 'to dominate, subject' and sui ^  / -^r (  pacify.' One can perhaps interpret that, when the bone graph for gi  ) 'to subdue, was created, it  was intended to show a woman being subjected (or perhaps 'taken as wife'). On the other hand, the present graph is unusual in that in most instances the 'woman' element is standing rather than kneeling. This would seem to support the idea that a gi jj^ woman who was equal to a rin  \  was a  (which means 'man' by default).  Usage As noted above, probably in its primary meaning 'wife,' perhaps specifically 'principal wife.' Since the OBI and bronze usage of shuang  ^  (II. 1 .ii. 186-191) to  refer to ancestors' wives does not survive in soft texts, one can only rely on the inscriptional context to determine the difference in meaning between it and gi ^fr . The "te  first thing one notices, is that shuang y$  occurs far more frequentiy than pj.  (all  variants: over 200 instances versus about 25 instances respectively ). It is also possible 1  that shuang  ^  was a term reserved for dead people.  There are no examples of it  applying to a living person in OBI, but there is one clear example of gi referring to a living person:  I have given approximate figures due to the uncertainty over which graphs should be recognized as variants, but the ratio is perfectly clear. 1  189  (Tongzuan 431) The king prognosticated, saying: "There is a curse. Perhaps there will be a trouble-report coming." When it reached the ninth day, xinmao, there was indeed a trouble-report came, from north Y o u . Lady (?) Zhu (?) reported, 2  saying: "The Tu tribe abducted (from) our fields ten people."  50  ^  JGWB 4423 (unidentified)  S.12.3x2  JWJ 9.2855: xu. 4)  , "S &  SW9a.7b:^f  Als  . XX ¥  S W D Z 9a. 18a: Now written $|  "T  Tr: Xvi ^  .  , XX 4  . fi  ...  [Duan also emends the definition to yi xia m £ o  'the hair under the chin.'] means 'facial hair.'  It consists of xi£\ "|T  'head' and shan  ^  'ornamental hair.'  Analysis *For the identification of ^ distinguished from mk ^ this error out to me. 2  For vou  3 5  as /  , interpreted as a i n . ^f'^. , see JWJ 2.335. It should be (JWJ 3.1081). I am grateful to Professor Takashima for pointing  a place name, cf. Tongzuan 513 (S.93.2)  190  The graph depicts a person with a beard.  Usage Seems to be a person's name.  191  Part 8:  The graph \  \  , Guo Moruo has suggested (see below), depicts a spoon with a  hook at the back for hanging it over the rim of a cauldron. As an independent graph in OBI it is always used as a rebus for the word now written  , with the meaning 'female  ancestor of more than one generation back,' i.e. grandmother and earlier (for 'mother,' mu is used). I therefore translate the tide as 'ancestress.' As a component in other graphs, it is used both as 'spoon' and as 'female.' The female usage is restricted to animal graphs, in which it is used to denote the female of the animal. In other graphs it simply represents a spoon, as in zhi t=j  / O  'spoon' +  'mouth' = 'delicious.' Since the present element depicts a spoon, it is not actually a human figure element, but I have included it in my thesis in order to clear up the confusion that surrounds it. After discussing the graph itself, I shall divide the present part into three sections according to the role of the element thus:  i. Female ii. Spoon iii. Phonetic (?)  192  2 ^  JGWB 1024: bj fc_ ; 1421: bj  -kt  S.5.2xl0 (excluding examples followed by tiangan  JWJ  8.2679: bj  SW  8a. 15a:  Tr: Bj /\  t  "C. means 'to line up in a row.' It is the character rin A.  ] turned round. Another meaning of bj  food,' also called a si  b  'person' [seal form:  is 'that with which one scoops up  'spoon.'  Analysis The  graph depicts a spoon, so it is the second definition that SW gives that is the  primary meaning. This explains its role as signific in the character chi ^Sr 'spoon' (the shi Guo  element is phonetic, but at the same time may originally have depicted a spoon, as Moruo maintains, so the bj  augmentation).  b  signific could simply be a later elucidatory  Its use as 'spoon' is also attested, albeit sparsely, in the classics, there  being one example in the Yijing and one in the Shijing:  ^ "And  "  -xJL t_  ^  . (HY 31/51)  he does not let fall the sacrificial spoon and chalice."  1  /M- ^-  (HY4g/203/l)  "Long and curved are the spoons of thorn-wood."  iTranslation from Wilhelm 1967:50.  193  According to Guo Moruo, the 'arm' near the top of the graph represents a hook that served to hang the spoon over the rim of a cauldron, and that exacdy this situation can be seen in  (O/NJWB Fulu 2.34a.2/2.303)  It may also be seen in the bone graph  (O/NJWB Fulu 2.34a.3/2.304)  0  jjfg  (Yicun 895, JGWB 333:  *  )•  Curiously though, none of the Shang spoons that I have seen in publications have such a hook, so the claim that  £  depicts a spoon with a hook still awaits archaeological  corroboration. Although the bone graphs for bi t i  and r i n A_ are often quite similar, there  are definite differences, and in the bronze script and all subsequent styles of script they are quite distinct. I think the confusion between the bone forms is largely due to inaccurate transcriptions that tend to homogenize the differences.  Usage It is used for the word now written  ancestress.  194  Section i: 'Female'  195  1347  j  JGWB 80: pin  $t  S.213.4x53  JWJ 2.303: pin  4t  SW2a.3a:W  $ 0  : £  $t  * £  .  S W D Z 2a.5b: Pin ^-L. refers to the female of all domestic animals. T n P i n ^ t . means 'the female of domestic animals.' It consists of niu -4" and hi t  'cow'signific  phonetic. The Yijing says: "Care of me cow brings good fortune."  1  Analysis JGWB 80 identifies all the 'animal+ b ' graphs as p i n $t.,  and the SW  definition provides the justification for this. Cf. the SW (ibid.)definition of mu -zr  chiifu  v  |?  as  /--.  X - 'the male of domestic animals,' which leads JGWB (78) to identify all the  'animal+shi dr ' graphs as m i  . L i Xiaoding is more cautious, and identifies only  the present graph as pin The reason why Shima lists the 'animal+ b. ' graphs under r6n is that he does not distinguish hi fc- from r£n A  in his index,  as graph components.  The two  graphs are generally distinct, but Shima's standardized orthography tends to blur the difference.  What we really do appear to find confusion between here is hi \  'spoon' and dao  I) I 77 Tcnife.' Perhaps  here is really  t  I, written upside down.  Another possibility is that we are dealing with two different graphemes: pin and * i f lj , perhaps indicating 'bullock.' Note that * ht  /  (S.220.1) and * M  cow, (S.220.2)  are clearly differentiated by context, * ^ t . occurring as a noun referring to a sacrificial victim (hence probably 'sow'), and *^-'J 'to butcher.' However, the use of* 4'1  occurring as a verb, perhaps with the meaning a n d * | l j appears to be identical with that of pin  !HY 19/30. Translation from Wilhelm 1967:119.  196  #t  and*j£t  respectively, and Shima does not seperate them. JGWB has the remark  under its first example:  t . "H  %  'hi t  U  is mis-written as dao 77 .' It  seems that we have to accept this explanation. Pin ^-t mu  and the other 'animal + t  ' graphs are fully analysed under JWJ 2.291  . However, it is attempting to analyse them according to the same principle that  has led people astray. Luo Zhenyu is on the right track (JWJ 2.291) when he says that nyi does not contain  phonetic as SW (2a.3a) maintains (which is a pretty awful  phonetic anyway), but shi analysis, that shi  'male' signific. However, he errs in accepting the SW  consists of shi  "f" 'ten' and yi. —•  2.296) gets nearer the truth when he says that  1  and f  reproductive organs, but mistakenly identifies them as liao The identifications as shi  "JZ and  'one.'  M a Xulun (JWJ  depict the male and female 3  and y_£  respectively.  are in fact correct, but they do not operate  according to the same principle: they both represent 'male' and 'female,' but whereas -i.  actually depicts a male organ, \  depicts a spoon and only indicates 'female'  through its homophony with the word now written  This is an example of what I call  a 'rebus signific,' and they are very rare. We come now to the problem of whether to recognize all the 'animal + t graphs as pin ifffc. and all the 'animal + -j: ' graphs as mji at  1  , as JGWB does. Luo  Zhenyu (quoted by Yang Shuda, JWJ 2.294), agrees with JGWB's position, and maintains that you %, 'doe' is also a variant of pin at  , and the pronunciation you was  a later invention. Yang Shuda however seizes on you ^  as a relic of a bygone age  when Chinese had separate words for the male and female of all domestic and familiar animals, just as English does. He proceeds to identify as many of the bone graphs in question with characters in the Erya thus:  197  I%  = jia | | _ 'stag'  ^  (no character )'boar'  /£fc  1  (nograph)/J± = zhif<|  /Jg  'stallion'  $f  = you. = h a  'doe'  frt  /I* =sM 1  'sow' 'mare'  The problem with this, as L i Xiaoding (JWJ 2.298) points out, is that these identifications are based purely on meaning, so there is no way of knowing if these really are the words that the bone graphs represent. He notes that the character sh£ SW(4a.l6b) defines  fen  as zang jff  , and zang Jff  4a.34a) changes the definition of fen  is not in SW, and that  as 'ram.' Duan Yucai (SWDZ  to 'ram' and zang jf^  to 'ewe' in order to  make SW consistent with the Erya, but this is quite arbitrary. Words for the male and female of animals are bound to differ between dialects and across time, so Li's caution in leaving these graphs unidentified and simply transcribing them  etc. must be followed.  Usage As a sacrificial animal.  1348  JGWB 2244:  tewen  y L of pin Jft and nul Jf±  S.214.1xl  Not in JWJ  Analysis  ^ u t in SW we have the characters x i i  (9b.l6b) and jia  (9b.l5a).  198  This curious graph only occurs once, and appears to contain the symbols for both 'male' and 'female.  1  Perhaps in this instance the |  element really is dao 77  'knife'  and the graph thus indicates 'bullock.' Another possibility is that the Shang herds did actually produce a hermaphrodite freak-this would explain the rarity of the graph (though the inscription says 'two  ,' and it seems unlikely that there would have been two at  the same time; it also seems unlikely that the Shang would have had a word for 'hermaphrodite cow' to correspond to the graph  1379  ).  JGWB 80: pin ift.  S.217.3x2  JWJ 10.3127: 3jt  (not in SW) [this reference not given in S]  Analysis The graph consists of quan |  /j£  'dog' and the sign for 'female,' and thus  probably means 'bitch.'  U s a g e Obscure. In Houbian 2.5.10 (the graph given above) it is a person's name.  1362 S.215.3xl5  JGWB 80: pin  #t  199  JWJ4.1345: £ t (notinSW)  Analysis The graph consists of yang cases it looks more like dao  % 'sheep' and the sign for female, though in some  i) I T} 'knife,' so we have the problem of deciding  whether it means 'ewe' or 'wether.'  Usage As a sacrificial animal.  1405  JGWB 80: (including jfy  7 $  ) pin  #t  S.220.1x6  JWJ 9.2989: ^ j  = frt (not in SW)  Analysis The graph consists of shi 4§C 'pig' and the sign for 'female.' Xiaoding only identifies the graph Lan's identification as ^'J the case of pin it  which clearly contain bj t  (S.220.2) as fat , even though he quotes Tang  (also not in SW). L i gives no examples of  and %t  . Although in  it is hard to see any difference in usage between the forms and those which apparently contain dao 77  case a difference in usage is detectable. While the form with bi b the form with dao 77  Curiously, L i  , in the present  is a sacrificial victim,  may be interpreted as a method of disposal:  (Zhuihe 59, ap. S.220.2)  200  Gengshen-day cracking: Issue the call to fetch the butchers and chu-soldiers.  J  : fa \ %9L "f {  #L  .  Z±  (Yibian 2833)  Tested: Butcher piglet to Father Y i .  However, due to the elliptical nature of OBI, the second example could also be interpreted as '(offer) female piglet to Father Yi.'  Usage As a sacrificial animal.  1440^  (Xubian 5.26.8)  J G W B 80 s% (Qianbian 6.46.6): pin %\_  S.222.4x1  JWJ 10.3050 (same example as J G W B ) : . f t (not in SW)  Analysis The graph consists of ma  ^  'horse' and the sign for 'female.'  Curiously,  JGWB and JWJ only give the graph that Shima lists after the present one, which could be interpreted as containing 'knife.' The usage of both graphs is too fragmentary for any contextual corroboration.  1563  JGWB 497: %L  (not in SW)  201  S.236.3x3  JWJ 4.1291: %JL (not in SW)  Analysis The graph consists of zhui i-t identifies it as cj  'bird' and the sign for 'female.' Y u Yongliang  filM-'female of birds,' but L i Xiaoding sticks to his agnostic position.  Usage Place name.  202  Section ii: 'Spoon'  203  15  fc=J  JGWB 602: zh£ a  S.6.3x43  JWJ 5.1643: zhi  SW5a.l4b: § Tr: Z h i  ||  , |£  .  A l -fr ,  b  <f  . R  Q means 'delicious.' It consists of gan  ... 8  :  4 i  'sweet' signific and M  phonetic... 1^ is the old form of zhi E3 • [The old form appears to contain qian 4" 'thousand,' i.e. 'thousand' + 'sweet' = 'delicious.]  Analysis The graph consists of b i  t.  'spoon' on top of k o u  r2  'mouth,' thus  representing the idea 'fine tasting.' B£ t_ is not primarily phonetic but, although the initial is very different, as far as one may tell from the E M C value p- (the E M C initial of zhi O  tp- is probably palatalized from *k-, cf. ji  could act as a phonetic hint. Note how zhi e  kej), it is in the same O C rhyme, so it functions as signific in ch£ng  %  'to  taste' (the phonetic is shang fa) ). In the seal graph, the mouth element has been changed into gan L i Xiaoding fails to distinguish zhao different, having dao ^ zhi  O  I  from zhi d  / ~P on top as phonetic. Their usage is also clearly different:  is the name of a Period I military leader, while zhao ^  name of a Period I tribe ^ •77  //  occurs mostly in the  'the Zhao tribe.' L i (JWJ 2.357) only recognizes  (S.359.1) as zhao e. , though he recognizes both t J and ^ /  , though it is clearly  as zhao  "77  ^  in bronzes  (as does also Rong Geng-see O/NJWB 112/135). The latter, more complex graph occurs in OBI only as a Period IV place name.. This graph is thus separated in time from both zhi and zhao r?- . It is significant that it occurs right at the end of the Shang dynasty, since at the beginning of the Zhou dynasty, this was the place where Duke Shi of Kang  204  )%  ^jik.  n a £  *  m  s  ^ ^' e  a n <  * * ^ l a  s o o  c  c  u  r  s  a  s  place name in early bronzes. This  a  character is not in SW, but in soft texts what is commonly acknowledged to be the same place is referred to by the character e . , read shao. and it seems at first sight that * 8> /  contains zhao ^  I 77 ,  as phonetic. In some bronze forms, we clearly do have djo f  but in other forms it is written  , which looks like rgn A.  be due to confusion with the descending hand elements ^ ^  , or  that flank it. In the bone form  than dao.  (JGWB 92), it sometimes looks more like  3 , which appears to  h / 77 . I think on  the whole it was intended to be dag X) , but it is not always well written. In some bone variants, the zhao ^  phonetic element is lacking (S.359.2). Some forms appear to show  two hands placing (or lifting up?) a wine vessel on a stand. It is not known what meaning this graph was originally designed to represent.  Usage Occurs mostly as the name of a military leader. There are no examples of it in its primary meaning.  2487  B  JGWB 3868 (unidentified)  S.385.4x7  JWJ 3.855: geng  f  SW 3b.5b: [ ' § ,  *« f .  JjlSij | \  3-  :f «  %  . AA |  .§ :£  XL h  ,  £  .f  . #  4- f  0  :  *  %  205  Tr: Geng  means 'five flavours harmonized soup.' It consists of l i ^  vessel'and gao  lamb.' The Shijing says: "There is also the well-seasoned soup."  1  It is sometimes abbreviated as ^ 'delicious' and geng gas  'cooking  It is also written J $ j , consisting of mei | ^  .  abbreviated. The small seal form is written  lamb' and mei  , consisting of  'delicious.'  Analysis L i Xiaoding correctly analyses the graph as consisiting of r&u. | ^ 'spoon,' min  'meat,' M t l  'vessel,' and little dots representing gravy. He accepts Luo Zhenyu's  identification as geng | ^ , although really there is no cogent basis for doing so.  Usage Context fragmentary, but probably some sort of sacrifice.  2520 Jg>  JGWB 333: s i  S.388.4xl  JWJ 3.854: g  (notinSW)  S W 3 b . 6 a : f | l , ft  f  S W D Z 3b.l2a: [Emends:]  ft*.  Ht  {fit f  *HY 81/302. Written  2  $  in the present text.  «.  A-  t&  . »*  . If  %  «  :  4*  *  $  .^  , $4  f  ,  «)  206  Tr: Su  means [the contents of the cauldron. The Shijing says: "What were the  vegetables? Bamboo sprouts and reed shoots." ] In Chenliu . jian 4^. 1  called si  . It consists of li  also written with shi  %  2  'cooking vessel' signific and si $J».  'rice gruel' is phonetic, ffc* :  'food' signific and shii i %~ phonetic.  Analysis The graph occurs without context, so it is hard to be sure of the proposed identification, but I think it is at least clear that it contains hi t 'person.' It seems likely that the dong $ lido  / ^  'spoon' and not rgn  k  / j ^ . element is phonetic, while the apparent 3  element probably represents the contents of the cooking vessel.  H Y 71/261/3. I am grateful to Professor Takashima for informing me that this was a place name.  X  2  3  Dong $  towrj < *tat] andfiM jjf / w  cuawk < *tak (initial uncertain) are both graphically w  and phonetically similar (being in corresponding nasal-final and stop-final rhymes). Shit ^ show a bag tied at both ends, and Chinese scholars agree that dong $ 6.2029).  seems to  shows something similar (JWJ  207  Section iii: Phonetic (?)  208  2652  fc^  JGWB 1668: ^  (notinSW)  S.406.2xl87  JWJ 7.2578:'^  (notinSW)  Analysis The graph consists of bj ^  /  'a hand-held net' with bj fc- on top. B j t  (pj i* < *psj ?) is phonetically very close to bi ^ they are in, zhj ^  &  and zhi ^  graph is a variant of hi Jp.  (pj it < *pac). The two rhymes that  , are very closely related. I would guess that the present  , with hi t l  added on top both as a phonetic hint and also to  suggest the thing being caught in the net. This does not mean that a spoon is being caught, but the fact that the phonetic element is placed in the mouth of the net is probably intended to be suggestive. Compare huli ^) / \ % (III.2.376), in which ge_  'spear' is  phonetic, but at the same time is placed suggestively in the hands of the offering figure-it does not necessarily mean that the graph refers to the offering up of spears.  Usage Occurs principally as the name of a military leader. This contrasts with the form , which usually refers to the capturing of game (S.404.1-406.2).  209  S U M M A R Y OF CHAPTER I  Since my chapter on the rin  *\  element is the most complex chapter in the thesis,  it might be a good idea to summarize the main points before proceeding to Chapters II and IH, so that the reader may have these points in mind when considering how the use of the dl  A  and \  elements differ from the of the r i a  To start off, r i n \  f elemenL  is an 'all-purpose' element that could be used in graphs  denoting words that had anything to do with humans. In order to bring some organization to my material, I divided it up into six divisions:  i.a Type of human i. b Specific human ii. Human actions iii. Body parts iv. Miscellaneous v. Phonetic  I believe it was useful to do this, because it is not necessarily the case that in a script of the same nature as the oracle bone script, a single human element should be used to cover all these areas. For example, the designers of the script could have chosen to devise a separate grapheme for making up graphs that referred to types of humans, or a special grapheme for making up graphs for people's names. It was therefore necessary for me to attempt to show that the rin \  element does indeed have all these uses. This classification is also  210  heuristic as, for example, one might have avoided interpreting graphs containing the rin ^  element as people's names if one did not have reason to think that this was merely a  standard usage of this element. Or one might have felt that such graphs were only being 'used as' people's names and really had some other meaning, which one might then endeavour to find out and be forced to say 'meaning unknown,' when in fact it would be quite acceptable to conclude that the graph had no other meaning and was designed to represent a person's name in the first place. I then identified a number of variants of the r i n element: \\  ,  j  , |  , and \  ,  clarify a point of confusion).  ^|  ^  people, sometimes standing for 'many.' is rin  \  = *  ,  shows the emphasis on the abdomen, and is used in = * —> represents a person lying down, and  is used in graphs to do with sickness, sleeping and dreaming. h\  ^  ^  (this last not strictly a variant, but included in order to  graphs relating to that area of the body. M  fashion).  ,  upside down.  } ^  shows a couple of  shows a person sitting (as in the Western shows the emphasis on the hair. And  \  depicts a kind of spoon but is used as a rebus for the word M 'ancestress,' now written •j&L . The conclusion that rin.  ^  sometimes represents a person lying down (when it  does so can only be determined by the analysis of particular graphs) is something that I would especially like to draw attention to, since it is hot otherwise obvious. Some people may feel skeptical about this, but I have tried to show that it makes sense. In contrast to the several forms and uses of the r i n uses of the da ^  and  \  element, the forms and  elements that I examine in the remaining two chapters can be  described much more specifically.  211  Chapter IL  Parti:  What deterrnines the use of the d l  A  A  ^  element is perhaps the most difficult to  work out. At the heart of the matter is the problem over its primary meaning. Is it really intended to represent the concept 'big,' or is this a rebus usage? Since the word d l  is  not known to have any other meaning apart from 'big,' it seems most likely that it is indeed intended to represent this concept.  But how does it represent it? The usual  explanation is that it represents a man standing with his arms held out, drawing himself up to his full stature, and thus implying the idea 'big.' I would like to modify this slightly, and say that d l  is intended to depict an adult. In Chinese, daren  , literally  'big person,' seldom refers to physical size, but either means 'great (i.e. important) person' or 'adult.'  1  The graph d l ^  then may be interpreted as portraying a person (male by  default) who has reached the age where he is able to stand firm and confront his responsibilities as an adult. One is reminded of Confucius's saying: Support for the idea that da ^  j=L "\~ \$p JA .  is intended to depict an adult comes from the graph fu.  'adult male.' It represents, in a stylized fashion, a man wearing a cap that is pinned onto the hair. This is the cap of manhood that was ceremonially  ^ e e Dai Kan-Wa Jiten 3.414-5.  bestowed,  212  according to the ritual texts, at the age of twenty, and signified that a male had reached the 1  age where he assumed his various social responsibilities as an adult (e.g. the duty to get 2  married). One may compare the fact that the graph for gr \  element. We see from this that the r i n  significance [+ADULT], whereas the d l  ^  the graph for 'child' to be written *  .  ^  ^ / ^ » 'child' contains the rin  element did not carry the inherent  element did. Thus it would be impossible for  This helps to explain then why there are no clear examples of d l  ^  as a graph  element in OBI having the significance 'big.' In most cases it simply represents a person. Its use is usually conditioned by the desire to depict both sides of the body (i.e. in concepts where the involvement of both legs and/or both arms was felt to be important to the depiction of the concept), and is thus used instead of rin  ^  when the concept in question  was felt to be most easily or most naturally depicted from the front. This makes it particularly suitable for use in graphs designed to capture the concept of standing still, where it was felt desirable to portray both legs in order to emphasize the idea of being stationary.  This use comes out particularly when one compares it with the r i n  \  element, as I have already done in the Introduction. However, where the idea of subjection was felt to be more important, then the kneeling figure was used, even when it was desired to depict both arms, as in the elements ^  ,  ,  ^  and J*"  which I examine in the  next chapter (Parts 2, 3, 4 and 5).  B.g. Liji.Neize: ^ ~T~ 7&> "At twenty, he was capped" (Legge 1885.3:478 1J34). The Yili and Liji both have chapters devoted to this ceremony (Shiguanli and Guanyi respectively), but in neither chapter is the age of capping mentioned. Steele (1917.1:266) says that the traditional ages for the capping of boys and the pinning of girls should probably be regarded as inferior limits. I say 'duty,' since the main purpose of marriage was to provide descendants to maintain the sacrifices to the ancestors. Steele (1917.1:266) notes that a man was expected to be married by the age of thirty. l  2  Zhouli.Diguan.Situ xia.meishi says:  ... ^  ' ~- "f*  vfp ^  ,  *f~  (Lin Yin 1974:144) "Officier des mariages (Me'i-chi)...il ordonne que rhomme a trente ans prenne femme, que la fille a vingt ans soit mariee" (Biot 1851.1:307). The Baihutong says that marriage is not an occasion for congratulation, because it signifies that the son is taking over from his father as the provider of descendants (see Tjan 1949:2491|241). Thus marriage is clearly portrayed as a duty.  213  After dealing with the graph d l ^  itself, I shall divide the present part into the  following sections, according to the role that I deem this element to play in the graphs that I assign to those sections:  i.  Graphs to do with standing  ii. Other graphs involving both sides of the body, or most easily depicted from the front iii. Miscellaneous  214  153  j\  JGWB 1239: da 7v  S.28.2x259 (excluding many collocations)  JWJ 10.3199: d l  7v  SW 10b.2b:  SWDZ 10b.4b: This definition is based on a sentence in the Laozi. [ch.25], which actually reads:  "Hence the way is great;  heaven is great; earth is great; and the king is also great."  1  Tr: Da 7v : Heaven is great, earth is great, and man is also great. This is why d l -fc depicts a man. [ i l ] is the old form of d l  Analysis Karlgren (GSR 317a) says: "The graph is a drawing of a man (grown-up person)." I think this is probably correct, and I have already explained how this may be related to the concept 'big.' L i Xiaoding says that the meaning 'big' is a loan usage. He does not state clearly what he thinks the original meaning was, but presumably he thinks it meant 'man.' 'Big' is indeed the sort of concept that is very hard to portray pictographically. opposite, xiao  Its  * ' / ' J 'small,' consists of three short vertical lines, perhaps N  representing grains of sand, or perhaps just anything small in an abstract way.  Usage  In most cases it clearly means 'big,' and I believe this is its original meaning.  translation fromD.C. Lau 1982:39, |57.  215  Section i: Standing  216  165  JGWB 3849 (unidentified)  «»« S.30.3xl  JWJ 10.3209:  SW 5a.l7b:  (notinSW)  (f ,^  f i i ^ f j k . # & tl%k . U jt , £ &  Tr: I i j | | means a bell-support. It is decorated with ferocious beasts. The character consists of hu.  'tiger stripes,' with y i ^  depicting the legs underneath.  is also written with jin JE. 'metal' signific and j i  phonetic.  : ji : the seal  ii  form of j i J&L is abbreviated.  Analysis Shima transcribes the present graph with the three circles that occur underneath it in the inscription, but JGWB and JWJ both omit these circles, and indeed it seems better to regard them as a separate graph. Although this bone graph has not been identified, as far as I know, it is very similar to the bronze form of j i y^ji 'bell-stand,' so I think one may safely identify it thus.  4* (O/NJWB 772/630)  Karlgren gives  and jfjj  (GSR 78e and g) as variants of the same word, which he  defines as 'upright posts of a drum or bell frame' (as opposed to xun %) ~  mA  f  ~  the cross-beams), and this is the meaning it generally has in the bronze  217  inscriptions. The reason why X u Shen says it was decorated with fierce beasts, is in order to explain the role of the hu. yL  tiger-top element. I do not doubt that it may have had  such decoration (Duan Yucai 5a.43a quotes a passage from the Zhouli to support this ), but 1  the role of this element is of course primarily phonetic: h i < *ga?. Karlgren separates it under his pji )jl really just a sub-series of hu  . The ^  xo' < *xa?, j i  k'ta < *k'ar phonetic series, but this is  element underneath qu i ^ . is a form of qiu  'hill' which happens to have remained more faithful to the seal form Ml 8a. 16b), and is of course signific here, the primary meaning of pji Jit 'mound.' In the character j i  however, the  you can see, and never contained pju_ examples (from the wf  gta"'  (SW  being 'hill' or  element is corrupted from  , as  . Since Karlgren quotes one of the bronze  ), it is strange that he does not bother to remark on this. I  suppose he just decided to gloss over this problem. The present graph clearly does not depict a bell-frame, so it seems to me that this must be a specialization of a more general basic meaning. In the Fangyan (Jiaojian 5/38/37), j i  is defined as 'a high table of the sort that one puts before a couch'  fli) JL> . . .  /iiH  ). I think the basic meaning was probably 'a  stand, a support.' Note the alternative form j # § < , which suggests an etymological relationship with j i ^jft  kia  h  < *ka-ys 'to depend on.' The signific element  then,  which has not been found independendy, may be interpreted as showing an adult with both feet planted firmly on the ground, and the 'tiger-top' phonetic is joined on top in such a way as to suggest that it is the person's head. Thus d l ^  is used, rather than rin ^  , because it was felt desirable to portray  both legs in order to represent a concept to do with 'standing,' and the feet are also drawn in to emphasize this.  ^rom the chapter Kaogongji.Ziren (see Lin Yin 1974:462). The passage describes the beasts in terms of categories, so no specific animals, such as tigers, are mentioned.  218  Usage Although in bronzes it occurs with the meaning 'bell-stand,' the bone graph occurs only as the name of a person, Z i Ju £  167  ^  Prince Ju.'  JGWB 1023:yi  %L  S.30.3x4 (excluding its usage as a diviner's name)  JWJ 8.2675: y i $}C  SW8a.l5a:^ ,  £  &  .  M  K. , 1 £  . £  ,  £  *  &  £  .  S W D Z 8a.39b: [Omits w^i t ^ L , which he says was inserted due to confusion with yj , thus giving the opposite definition. Duan says  yigX-  is often glossed as zhi jJb  in the classics.] Tr: Y i $X- 'doubtful' means 'uncertain.' It consists of hua  Y- 'to change' signific  [Duan explains that 'after things change, they become stable,' but I think X u Shen may have meant this in the sense of 'changeable'] and shi  phonetic.  #z. is the old form of  arrow.  SW 14b.l2b: *t  , §  A,.  S W D Z 14b.26b: Neither hui  JX*  ,jk,  \L,  ^- nor s M ^  The analysis should be XL %T , it  can be phonetic, but zhi j £ - could be.  % , jk. $  'certain' abbreviated, and zhi _lk phonetic.  1  'from zi  'child,' y i  |L  'Child' and 'certain' combine to give the  meaning. [Duan revised the definition of y i $X- to 'certain,' but X u Shen's original definition 'uncertain' would fit in better here.]  219  Tr: Y i $3t 'hesitant' means 'perplexed.' It consists of zi hua  ^  'to change,' and sM ^  'child,' zhi j h . 'to stop,'  phonetic.  Analysis Duan Yucai attempts very bravely to wrestle with X u Shen's uninsightful analyses. If only he could have seen how simple the original bone graph is! The graph simply depicts a person standing still, leaning on a stick. The primary meaning was probably 'stand still, stop>fix on, settle' (see GSR 956a), which is the meaning that y i  has in  the rusheng reading, while the meaning 'to doubt, hesitate' that it has in the pingsheng reading is perhaps etymologically derived from this ('stand stillVhesitate').  1  (Since  Mandarin has lost its rusheng, I am unable to indicate these different readings in the pinyin transcription.) Duan's attempt to use these different meanings to distinguish between the two SW characters y i $ £ . and y i  is thus futile. The simple character y i  is the  direct descendant of the present bone graph, but it occurs only in SW. In texts, the minimum form is always yi been corrupted into zi  . In this character, the 'hand holding a stick' element has  %~ 'child' in the seal form, while the additional signific zhi jt-  probably indicates the idea of 'stopping.' It is interesting that in the simple character y i Jfc. (1 the 'hand holding a stick' has been distorted into X u Shen's hua exacdy the same distortion took place in the character lao ^ If the primary meaning of yi  V- element, since  /  was 'to stand still,' then this would explain why  it was felt desirable to show both legs. On may compare the above mentioned graph l|o.  The Guangyun only gives yj 4^ a pingsheng reading, with the meanings 'uncertain, to fear, perplexed, to suspect.' However, there is some early rhyming evidence for a rusheng reading (Zhu Junsheng 1834:5.15a gives examples from the ShujingMongfan and the ChuciDazhao), and Lu Deming's glosses on the Shijing and the Yili provide evidence for such a reading with the meaning 'stand firm' (see Pan 1  Chonggui 1983.1:209). Note also that some graphs containing yi -ZJC as phonetic have a rusheng and a  non-rushengreading,e.g.yf JfiL n + , rjik,yi  rji ,n.ik,yi , | f rj4,rjik.  220  fj  , which shows an old person leaning on a walking stick, where there is no motivation  in the concept for not using ren \  .  Usage Name of a person, especially that of a Period II diviner (see Keightley 1978:195, table 6), so the Shang bronzes bearing this graph inside or under the y i izL cartouche (see O/NJWB Fulu 1.16/1.129) may perhaps be identified with this person or his clan. The graph ffi^ however (S.324.3), which has a chi A  road element added, and is included  by both JGWB and JWJ as a variant of the present graph, is used as a verb. L i Xiaoding says it is used in the meaning hu&iyf t i l  'suspect, doubt,' and I think he is probably  right:  | : -4 0 £ 0  yu  .  ... £ S © : ^  4&& %>  H .  il . (Qianbian 136.2)  Tested: Today perhaps... The king read the cracks and said: "(I) suspect this (crack) (awaits:) harbingers rain." That day it did indeed rain. Third month.  172  ^  S.31.4xl09  JWJ 10.3251: li J£  SW 10b.8a:  JGWB 1263: H JL  221  S W D Z 10b.20a: Z M £  has been changed from s M \$-  [defined at SW 8a.7a as JI  'to stand'] by a 'shallow person.'  Tr: L i £  'to stand' means zhu %. 'to stay.'  It consists of djj, A .  [representing a  person] standing on top of a line [representing the ground].  Analysis The graph depicts a person standing on the ground.  Usage The graph occurs mostly followed by an object, such as zhong xfy , s_hi ^  or rin  , shu  , so it is clearly being used as a transitive verb. It would make good sense  to read it as li >fj£ 'go and inspect'(GSR 520b). Thus li shu JL and inspect the millet (crop).' As for zhong ty  , shi JjJL and rin  would mean'go , these probably  refer to military units, so a causative interpretation 'set up, establish,' which is a common meaning of li £  in soft texts, would also make sense.  Once again, d l ^  is used because it was felt desirable to show both legs in order  to represent a concept to do with standing.  216 M - ff\  JGWB 1264: bing  jL  S.39.3x72  JWJ 10.3253: M n g  SW 10b.9a: Tr: Bing % . means bing ^  'to combine.' It consists of two H U- .  222  Analysis The graph shows two people standing side by side on the ground. Shima also includes three examples in which the ground is not present (Shiduo 1.416 and 2.76, Yicun 222).  JGWB 1262 includes these as variants of b i n  ftf  , but Shima lists other  inscriptions with similar wording in which the graph is written jti  with the ground  underneath, so his inclusion is undoubtedly correct. SW (8a. 16a) gives a graph ft ft as the 'old form' of bi  , which is curiously similar to the bone graph (\([  interesting in the light of the word family connection between bing  jllL  . This is  and M  fcb that  Professor Pulleyblank has informed me of.  Usage In many inscriptions it seems to be the name of a person, but there are also some inscriptions in which the interpretation 'together, to combine' would make sense, e.g.  (Houbian 2.9.1)  On the seventh day, jisi-night cutting into (gengwu-day)...there was a new 1  big star together with the Fire-star...  4  9  i  |f  t  ...  Today mix new wine [e.g. with y i  (Houbian 2.9.1)  'spiced wine'?].  Another interpretation of this graph proposed by Serruys (1974:106, n.35), is that it stands for ddu Iff]  1  (sic =l!frl ), which SW (3b .7b) defines as y i I S . 'to meet.' This meaning would fit the OBI usage between dates very well. If we understand 'meet' as 'intercept,' then this can still be related to the meaning 'cut.' My translations of these sentences otherwise follow L i Xiaoding's interpretations.  223  Section ii: 'Bilateral' and 'Frontal'  In the majority of graphs containing the d l  element, I think that its use was  determined by the feeling that the concept in question was most easily or most naturally depicted from the front. This is particularly the case in concepts where it is felt necessary to portray both the arms or both the legs, as the present section will show. The ren element cannot be used in such cases, because only one arm and one leg are apparent in this profile view. It is interesting to note here that, although the Egyptian script has a much greater variety of human figures in it, they are invariably depicted from the side. The human figure is never frontally depicted. This seems to be true also, in general, of their art. The Egyptians were quite happy representing both arms and both legs from the side, by adjusting their relative positions so that the near limb did not hide the farther. In the Shang script there are one or two graphs in which both arms are portrayed from the side (e.g. zhi %  I  , showing a prisoner with both hands bound in a sort of manacle), but  there are no graphs in which both legs are portrayed from the side. This naturally supports element is largely structurally conditioned.  224  160 ^  .^  JGWB 1250: zg  ^  S.479.4x9  JWJ 10.3213: m  SW10b.3b:  ^  ^  , 4f  A  . M  ^  , | L  . K/ . . .  Tr: Zg, "^L means 'to incline the head.' It is based on d l  , and is a pictograph  [showing the head inclined].  Analysis The graph shows a person with the head inclined to one side. Thus the primary meaning must have been, as X u Shen has it, to incline the head. It is now usually written , and SW (9b.9a ) has this as a separate character, with the definition ceqing 1  'to incline.' I do not know what the origin of the form fft  does not seem to be derivable from the bone form ^  is-the seal form  . The word ze J£  now  seems to be restricted to its specialized use in poetry to refer to the deflected tones. The word zg JJ^ 'the sun slanting towards the West,' is another specialized usage. I assume that the d l ^  element is used in this graph because the designers of the  script felt that inclining the head to one side was most easily depicted from the front.  Usage Apart from a few obscure exceptions, it occurs mainly in the name Wang Ze S~ which, being the beneficiary of sacrifices, must be the name of an ancestor. There do not seem to be any examples of it in its original meaning.  *SW gives |X  ^  m  e  zhouwin form, and says that the z£ jk> element is both signific and phonetic.  225  168  ^  JGWB 3002 (unidentified)  S.30.3x30  JWJ 5.1825: yang JJ^  SW 5b.l0b: |  J ^ i ^ . i A ^ t n i ^  .  , JL  J L > .  SWDZ 5b.26b: [Explains that 'middle' and 'side' depend on each other for their meaning, i.e. in order for something to be in the middle, there has to be something on either side. This is typical of the Chinese 'relative' way of looking at things] Tr: Yang *fe  means'middle.'  1  It consists of da  D | ;7v here represents a person. The characters yang conception.  [SW la. lb defines pang  asbji^  inside jiong f") and pang ^ 'vast.'  'wilderness.' are similar in  X u shen's meaning is  unclear, but perhaps he is trying to make a connection between the 'edge' of a wilderness and the 'middle' of one.] One source defines it as long lasting.'  Analysis Ding Shan agrees with Dong Zuobin's description of this graph as depicting a person wearing a cangue, and says that it is the primary form of yang defined in SW (3b.4a) as jingzhi  * ^  jjtjS  , which is  'neck-strap.' Karlgren defines it as 'strap on  breast of horse or ox' (GSR 718f), and describes the bronze form as 'a man with a carrying-pole, supported in its centre,' but considering that none of the characters in this  1  Pulleyblank 1986 argues that yang ^  yi ^  'in.'  basically meant 'inside' rather than 'middle,' and is a derivative of  226  phonetic series relates to this meaning, I do not know how he can be so dogmatic. Ding Shan notes that the Zheng Xuan commentary to the occurrence of yang % ^ in the Shijing (HY 50/205/5) defines it as M  ffi  chariot' (  'to carry,' and continues 'it is like a horse 'carrying' a ). However, as Karlgren (Gloss 646) notes, the  earlier commentary by Mao treats the expression yangzhang binome meaning shir6ng  ^  in this ode as a  'disconcerted, perplexed,' and Karlgren identifies its  components with the words yang  and chang TH»), which occur separately with this  meaning in other early texts. Since the expression rhymes, it does indeed seem better to treat it as a rhyming binome, and the Mao commentary, being earlier than the Zheng Xuan commentary, has greater authority. It seems better simply to accept that yang  means  'horse's neck-strap.' The bone graph probably depicts a person wearing some kind of collar round the neck, the frontal view making this easier to represent, and the application to a horse's neck-strap would be a later development. One final point is that, although the graph probably depicts a person wearing some kind of collar, one cannot help being struck by its 'centrality.' I think this is a conscious part of the way in which the graph was designed.  Usage Person's name, mostly in the name Zi Yang %~ <$JL  175  ^  S.32.4xl97  JWJ 10.3211: yi  JGWB 1249: yi  #  'Prince Yang.'  227  SW 10b.3b: $  , A  %  fa  Jhj . rl  A  , 4  &  fa  *~  •R  ...  SWDZ 10b.7b: Y i fa came to mean 'in addition' because the armpits are 'added' to the side of the body. Tr: Y i  fa  refers to a person's armpits. It consists of d l !7\. with the two armpits  depicted. [Xu Xuan et al: Now erroneously written  .]  Analysis The graph shows a person with marks indicating the armpits. form of y_i  It is the primary  'armpit.' Because it was used as a phonetic loan for a common word  meaning 'also,' the word for 'armpit' came to be written with the phonetic compound , consisting of rim $J turn consists of xi $  'flesh' signific and y i  phonetic. Y i 7^  'night' in  'evening' and y l fa* phonetic, according to'SW (7a. 10b). It is  not found in OBI, but the bronze form is ^  ( 3 & J^- , O/NJWB 916/1132). The  'evening' element replaces the right armpit. Dl  A  is used because there is an armpit on both sides of the body, and it was felt  desirable to indicate both armpits, as they form a natural pair.  Usage Mostly used in the sense of 'also,' or perhaps 'again.' I suspect that the graph ^  was invented specifically as a rebus for the word 'also' and was never actually used  to write the word 'armpit.' This may be another reason why the complex character y i was created to write the word 'armpit.' Whether or not this suggestion makes sense depends on the original purpose of the script. It is true that the oracle bone language is largely religious in nature, so one might not necessarily expect to see armpits referred to (unless the king had a sick one), but we do not know what other genres of writing the Shang had. Supposing that they came across the need to write the word 'also' before they had written about armpits? It would be perfectly natural, using the rebus principle, to  228  depict an armpit to represent this word, without having previously used the graph to mean 'armpit.' The same may have happened with any graph that is commonly used as a rebus.  182 ^  JGWB 1240: jia &  'place name'  S.34.2x8  JWJ 10.3201: jia  SW 10b.2b: $ Tr: Jja  &  , #  A> . kk  7v  4%  X  means 'to hold.' It consists of d l 7 v  k  .  [representing a person] embracing two  people.  Analysis The graph depicts a person apparently embracing two other people beneath the arms. Some of the meanings Karlgren gives are 'be on both sides of, support, press between' (GSR 630a), and these seem to be the sort of concepts that the graph is intended to depict. The word x i l  , which X u Shen uses paronomastically in his definition,  1  has very similar meanings: 'grasp, hold, clasp under arm, encompass, embrace' (GSR 6301). There is probably an etymological relationship here.  l  M  the ji  Ts fa  k e p < * k r a p . xia r  j  ~gep < *g ap (for the classification as y£ Hp. 3  rhyme, rather than  which the E M C values would seem to point to, see GSR 630 and Wang L i 1958:66-67). It  seems unlikely that £hi  dr + < *dr 8-y in the definition is intended to be paronomastic.  229  Place name.  186 187  $  188  JGWB 1242:  'overlooked by SW'  JGWB 1246: H  (not in SW)  189 191 tflw 190  JWJ 4.1161: ju  SW3b.20b:^. S W D Z 3b.44b: [Refers to meishuang the *-)(  $1  'the grey light of early dawn.' Explains  element as like the holes in a lattice window letting in chinks of light.]  Tr: Shuang $ | explanation] and d l  means ming 8 $  'bright.'  It consists of l i &  [see Duan's  'big' [i.e. implying great light].  Analysis A great deal of ink has been spilt over the identification of this character, and I do not have space to discuss all the identifications proposed here. L i Xiaoding enthuses over Zhang Zhenglang's identification as ju. ^  , but I find it far too complicated and contrived.  I think that, on graphic grounds, one has to accept Y u Xingwu's identification as shuang ^  . This creates a big problem over the meaning of the graph in OBI, which clearly has  nothing to do with the SW definition or any other known meaning of this character, but I think one simply has to accept that this problem exists, and some avenues of escape have  230  been proposed. Here are some graphs that O/NJWB 455/569 lists as bronze forms of shuang  The  Jj[  graph is clearly a direct descendant of the bone graph ^  . The elements  under the arms may be lamps, in which case the SW definition 'bright' could be its original meaning. The form on the  ^i.  simplified to criss-crosses. The  is very close to the modern form, with the lamps graph is used to refer to the wife of an ancestral  king, in exacdy the same formula as we find in OBI. The d l ^  element is used in order to give the graph a symmetrical design.  Usage Although there is much dispute about the identification of this graph, there is none over its meaning, which is perfectly clear. It usually refers to the wives of the ancestral kings, and may perhaps be translated as 'queen' (except in a few cases where it refers to the wife of an ancestral minister). Yu Xingwu suggests that shuang stands for xiang  staar)' < *srarj'  ?  staan. < *sans in the sense of 'helpmate.' Another possibility is h  that it may be connected somehow with the word shuang  staan. <*sran. 'widow.'  As Professor Pulleyblank has pointed out to me, if a queen outlived her husband she would be his widow.  192 S.35.4xlll  JGWB 708: wj± 2£  231  JWJ 6.2039: y/1  SW 6a.24b: ^  ; 5.1927: w l  , f  Ab . &  ^  **-  SWDZ 6a.67a: This meaning is now written Tr: W u $ h  means feng  .  or  ] «  tl  &  *  t  ,  .  jji. 'abundant.' It consists of lin  Some sources say that *)$t is the character m i and  . .[«  'forest' and tfj* .  'model.' It consists of d l  'big'  represents a big number [there is no such character in SW, but it may be  analysed as nian JT* 'twenty'doubled]. A forest is a lot of trees.  is conceptually  the same as shi jjfc 'many.' The Shangshu [= Shu.Hongfan] says: "All the plants are rich and luxuriant."  1  Analysis  s  Xu Shen is able to marshal so much evidence in favour of his analysis, one hardly feels charitable in gainsaying it. However, the bone graph simply depicts a person dancing, and is the primary form of wJt  'to dance.' The modern graph must be  descended from a later form in which the feet were drawn in, such as we find an example of in bronzes:  (NJWB 899, not at OJWB 730)  ^Tongjian 24/848. Translation from Karlgren 1950:33, |26. In the present text, the last two characters are written  IT  1&  232  This emphasizes the fact that dancing is mainly performed with the feet. In the bone graph, the distinguishing feature is the things hanging beneath the arms. Ritual texts suggest that they may represent plumes. L i Xiaoding quotes ZhouliDiguan.Wushi (Lin Yin 1974:126):  "Us enseignent la dance des plumes variees, et sont chefs de danse dans les cer6monies des temps de s6cheresse." (Biot 1851.1:269)  This describes one of the duties of the wushi ffi iff  'dance masters.' The Zheng Xuan  commentary to this, which L i also quotes, reads:  <TYN  r\>fc  afu W  .' Karlgren defines fy  'the hu&ng  (GSR 276n). The word huang ,2yu  'feather' signific and w£ng  i  s m a c  , $r  ^jx.  %  ,  i from splitting five-coloured feathers; it is also like e  as 'wand with silk pennons carried in ritual dances' is also written j£_ , which as you can see consists of phonetic. SW (4a.lib) defines this character as  'to dance with feathers covering ones head when worshipping the stars.' Note also that SW (5b. 15a) gives the old form of ^ i l signific and wang "t  as \JT , consisting of yjL %M  'feather'  phonetic. Thus one can see that the association between feathers  and dancing was very close in ancient China. In OBI, dancing is often performed to solicit rain.  It seems likely that y j .  'feather' and yjL $3  'rain' have always been  homophonous, or at least very similar in pronunciation (GSR 98 and 100 gives them the same reconstruction), so one wonders if there is some sympathetic magic involved here, or whether the plumes were simply a decoration. In spite of Xu Shen's definition of wji  as 1£  'to rejoice' (SW 5b. 15a), I  think it is clear that in ancient China dancing was mainly a ritual activity rather than  233  something done for pleasure.  1  It was an expression of harmony between the spiritual  world and the physical world. The emphasis therefore was on its orderly and symmetrical performance. The bone graph, representing the dancer from the front, with the plumes hanging down in a balanced pair, may be seen as symbolizing harmony through symmetry. The ritual nature of dancing is brought out clearly both in OBI and the classical ritual texts such as the Zhouli passages here quoted. Note also that the word yyji differs from wji ^  ?EL 'shaman'  only in tone, and could well be etymologically related, i.e. it could  mean literally something like 'dancer.' Ritual dancing was one of the main activities of the shamans, as one can see from the Zhouli.Chunguan.Siwu passage (Lin Yin 1974:269) that L i Xiaoding cites:  % i  < % S'i  *  %  f.  "Si le royaume 6prouve une grande secheresse, alors il se met a la tete des sorciers; et il appelle la pluie, en exdcutant des danses.' (Biot 1851.2:102)  This refers to one of the duties of the siwu  3iL 'chief shaman.'  An interesting contrast is formed by the Egyptian ideogram  1  ^ (Gardiner  1957:445, sign A32), which shows a man dancing from a half-frontal view (there are no fully frontal ideograms of humans in Egyptian), and occurs as signific in a word meaning 'jubilate.' There is no striving for symmetry in this graph, only for grace and balance. I suppose the Egyptians had a very different temperament from the Chinese.  Usage  though of course the two things are not necessarily mutually exclusive. The designers of the script may have had a bias towards the ritual context, as the Confucian texts certainly do, and this would distort our view of the nature of dancing in ancient China.  234  In bronzes, siA  is always used as a negative, but in OBI it is always used in  its original meaning 'to dance.' The equivalent negative in OBI is wang  200  X  S.37.4xl4  201  S.37.4xl6 JGWB 316: yi j |  202  ^  S.37.4xl  203 205  S.38.1xl t#  S.38.1x25  JWJ 3.803: yi | g  SW3a.21b: I Tr: Y i J I  JGWB 4694 (unidentified)  (= dH % )  , #  J & . / A ft"  , /A #  'to differentiate' means fsn.  ^  ; £  , *  'to divide.'  .  ft  ...  It consists of gong 7T"  [representing a pair of hands] and bi 7T . B i 7T means yji "j  'to give.'  Analysis The basic form of this graph shows a person carrying a basket on the head, and is 1  the primary form of djti. ^  taj < *tg-ys 'to carry on the head.' The primary form was h  borrowed to write a word meaning 'different.' The phonetic element zai  tsaj < *ts8-y  was then added to distinguish the original meaning. L i Xiaoding suggests that in the form (my 200), the top element is zi phonetic hint. In the form jr  1  Which I identify as gui £l£ $L  (seeIII.l.i.283).  ty  tst < *tsr8-y 'earthenware vessel,' and acts as a  (my 203) we can recognize c i i  I  dzsj < *&zkn as  (GSR 540a 'basket') in order to account for its phonetic role in gui  ^  235  a phonetic hint, and in the form ^  (my 205, which L i includes as a variant, though  many scholars are doubtful over its identification), we can recognize zi •? / ?T t s t ' < *tsa? phonetic. Thus the thing being carried incorporates a phonetic hint. This is typical of the design of the bone graphs. The graph is depicted from the front in order to show that the basket is being held on the head by both hands.  Usage Some of the above forms are used as personal names, but their usage is otherwise rather obscure. Takashima (1973:70 #18) suggests the tentative translation "I-influence," or otherwise "transform, change (?); protect (?)."  206 j | .  JGWB 1599: (including ?  ) gjn  f  S.38.1x49  JWJ 13.4013: flin %  S W 13b.l4b: |(x , 4£  Tr: Qin Jf.  j t  means 'clay.'  abbreviated.... 'H and X  A l J±. ,  rV . . . f f  It consists of t i _ £ ~ 'earth' and huang  ,  'yellow'  are both old forms of qin JL .  Analysis Karlgren correctly identifies the present graph as  , an element which does riot  occur independently (not even in SW), to which he ascribes the pronunciation jian on the  236  basis that it is the primary form of jian "ffi character gm jg_ til jt_  'distress, difficulty (GSR 480). 1  The  actually consists of the present graph as phonetic and, in the seal form,  'earth signific,' distorted from hui> k  'fire' signific in the variant bone form and  bronze form (the bronze forms show how the 'fire' element became distorted to 'earth'):  W In OBI,  (S.38.3) > ^  -L  j£  N  (O/NJWB 1715/2189)  the present graph clearly refers to some kind of calamity, so the meaning 'distress,  difficulty' is quite appropriate, and there is general agreement on this. However, there is disagreement as to how the graph represents this concept. Karlgren says that "the graph shows a man (prisoner?) with back-bound hands." L i Xiaoding on the other hand suggests tentatively that the hands are being held in front, in a respectful, attentive attitude, and that this is the primary form of jin £jf. 'attentive, cautious.' This would make its use for 'distress' a loan usage. In OBI, crossed hands usually indicates that the hands are tied (see m.5), so I think Karlgren is more likely to be closer to the truth, though other explanations are still possible. One might note, for example, that many of the characters containing this element as phonetic have to do with heat and dryness, e.g. han~han ^jl 144a), b i n  1  Note also jin  'dry, burn' (144b), and r i n ft  , variant of r i n $S  'to scorch' (GSR 'burn' (152i).  'die of starvation' (480o) and the homophonous ~®% 'famine' (480s).  I am tempted to see the bone graph as a person suffering from thirst or hunger. This would help to explain the addition of the 'mouth' element on top that we see in most forms. However, it is hard to reach a conclusive solution. Whatever the precise interpretation of this graph, clearly the distinguishing feature is the crossed arms, and this is most easily depicted from the front. (In the case of n i 'woman,' the arms are depicted crossed from the side, but this is due to the  1  The Guangyun also gives this character a qusheng reading.  237  greater desire to use the kneeling element in this graph, in order to suggest the idea of submissiveness.)  Usage It is used mainly as a 'disaster' graph, for which Takashima (1973:71 #19) suggests the tentative translation "dry; drought; cause drought," but the form  ^  also  occurs as an adjective describing sacrificial animals. Possibly it refers to their colour (e.g. 'clay-coloured'?).  210  ^  JGWB 719: chgng $ t  S.110.2xl39  JWJ5.1933:ching  SW 5b.l7b: #  £  , f  SWDZ 5b.45b: Ch6ng ^  &  .  M  X  , %$ &  .  J  #  [  X  means 'to overcome strength with weakness.'  Tr: Cheng ^ £ 'to get on top of means 'to enter'and jM Tjl . M  as 'hard black'].  . £  means 'to add on top.' Its use to mean 'ride (a chariot)' is one  aspect of this. E i ji£ A _  rji X  , g  ^  fit  'to overthrow, vanquish.' It consists of 1  here stands for x i i ?f» [defined at SW 10a.26a  In warfare, entering the hard black [= overcoming strength with  weakness] is called ch6ng ^  . ^  : the old form of ch6ng ^  'table.'  ^ee Karlgren GSR 1034m and Gloss 376 for this understanding.  contains j i  238  I find this analysis very hard to make sense of, and Duan's explanation does not seem to help much. Ji6xi& ^  %a. occurs as a binome in Honshu 91.4b, and Wilbur  (1943:281) translates it as "rascally and crafty" (referring to male slaves).  Analysis The graph shows a person who has climbed up into a tree. The original meaning must have been 'to mount, ascend' (GSR 895a), from which it became specialized to 'mount a chariot, ride,' which seems to be its commonest usage in soft texts, and in the qusheng reading sheng  it even has the specialized meaning 'chariot* or 'team of four  horses (as would be used to pull a chariot).' However, the meaning 'mount, ascend' does have some textual support, as in the Shijing line:  ^  4  J  t  (HY 32/154/7)  "quickly let us get up on the (house=) roof  Although we do not have any inscriptional examples of the character ji£ ^  , I think it  must originally have depicted a bird-perch. This is the meaning it has in the Shijing line 0.  -f  ^  "the fowls roost on their perches" (HY 14/66/2). Ch6ng  ^  shows  a person perching in a tree, so it is easy to see how these two characters could become similar in the seal form. The seal form of ch6ng  ^  must have evolved from a form in  which the feet were drawn in, such as we find in the bronze form:  %  (O/NJWB 738/907)  239  The depiction of the feet here seems to be intended to emphasize the idea that the person is standing in the tree, after having climbed it, but the basic meaning of the word that this graph denotes would seem to be 'to climb.'  Usage It occurs mainly in the name of the military leader Wang Cheng  ^  #L . There  H  'sun' signific  are no examples of it in its original meaning.  211  JGWB 809: z£ Si  S.39.1x24  JWJ 7.2187: zg  SW 7a.2b:  Tr: Z£ ; X refers to when the sun inclines in the west. It consists of ri a n d z £ JX phonetic. The Yijing says: "In the light of the setting sun."  1  Analysis Ye Yusen and Dong Zuobin analyse this graph as a semantic compound showing the sun and a person's shadow slanting. Whether  actually represents a shadow is a  moot point, and furthermore, as Professor Takashima has pointed out to me, the sun is underneath the person, when one might have expected it to be on top. It seems best then to follow the suggestion that Professor Takashima made to me, that ^ iHY 19/30/3. Translation from Wilhelm 1967:120.  may be a variant of  240  z£ £  I  (no. 160 in this section). Whereas the latter shows only the head tilted, the  former shows the whole body tilted. Further support for this analysis may be found on the bone Yibian 32 where, in the same inscription, we find the present graph written both and "Q".  In the second example, the  ^  B  element is completely on its side. Strangely  0 enough, the modern form ^  seems to be closer to the original bone form than the SW  seal form fj\B . The present graph is thus a phonetic compound, but it seems likely that it is the same word as z& j X 'tilt,' but was created especially to refer to the westering sun tilting towards the Hesperides. Usage Refers to a time of day, fixed by Dong Zuobin as 2-3pm.  223 ^  .  ^  S.40.2x6 224  \r\,  JGWB 1639 (both forms): Hng  S.40.2x2  JWJ 14.4131: ling  Analysis Luo Zhenyu identified this graph as ling f^t. in the sense of 'climb up,' and said that it depicted a person with one foot on the ground and the other on the rung of a ladder. As you can see, JGWB and JWJ accept this. However, it seems rather obvious that the graph shows a hand sawing a person's foot off with a saw (the second form, 224, lacks the hand). In some forms the remaining foot is drawn in fully in order to emphasize the  241  fact that the other foot is missing. The serrated line represents a saw, and is quite distinct from f£  |  / %  ( f  ), which represents a wall with steps up it. The bone graph from  which the modern element ling ^  comes is probably $  (see I.l.i.b.40). Therefore  the present graph may not be identified as ling I £ . . I do not know what modern character it should be identified with. It may have been replaced by a phonetic compound. One might suggest yue  "to amputate (as a punishment),' but this is not a very good  candidate for, as Professor Takashima has pointed out to me, yue occurs in OBI with this meaning, e.g. with y £ n g  ^-  already  'sheep' as object (as in the  inscriptions Yicun 153 and 404). In order to illustrate the fact that one foot has been sawn off, it is of course necessary to show both the legs. Thus this graph could only be written with d l  .  Usage L i Xiaoding says it seems to be a person's name, but the context is so limited that it is hard to reach any conclusion. It could perhaps be interpreted as 'amputate' in the following inscription:  j| : \K  ^  (ft)-  (Qianbian 6.55.5)  Tested: The foot-amputated convict will not die.  This could mean that the convict had just had his foot amputated and this had endangered his life, or he may have been amputated some time ago and the threat to his life was from some other source. On the other hand, l^ts  could be a person's name, and k6u 7& refer  to a convict that he owns. With no other context, the interpretation is very uncertain. The Duo Kou  9  7&> were used by the Shang as a military force (see I.l.ii. 1847), so these  people were of some value to the Shang state, and perhaps this explains why the diviners are here concerning themselves with the health of one of them.  242  235  JGWB 4705,4708 (both unidentified)  S.40.3x5  Not in JWJ  SW5a.l3a: $  , J|  SWDZ 5a.31a: K g " o f Tr: OJ  j& .  -  0  : ^  J$> • M  it , kk ^  .  is also phonetic.  'strange' means yi f |  even]. It consists of d l k.  'unusual.' One source defines it as 'odd' [opposite of  'big'and kg. °J  [defined at SW ibid, as kgn IT  'meat  sticking to bones'].  Analysis Kang Yin (1983:35) describes this graph as depicting a person riding a horse, and tentatively suggests the identification pi  , i.e. as the primary form of pi 05 ^ 'to ride a  horse.' Presumably he is implying that the lower element was later corrupted to kg  as  a phonetization. I think this identification is a distinct possibility. However, one should note that in none of the five instances of this graph is the element bestridden truly horselike, so I think rather we have to interpret the graph simply as a person stepping over an object, the exact nature of which is unclear. Most scholars believe that the Shang did not ride horses, but only used them to pull chariots (see Chang Kwang-chih 1968:237), so indeed one would not expect to see a graph that depicted a person astride a horse. The original meaning of the word pi would simply have been 'to bestride,' and only later would have come to be applied specifically to bestriding horses, and the use of pi 'odd' would be a phonetic loan.  <*]  to mean  243  If Kang Yin's identification is correct, then this solves several problems, the biggest of which is: how on earth does dl k a phonetic compound, in which only the dl relate the meaning of dl  +k£  n  2  mean 'odd'? If we regard it as  7^. element is signific, then it is still hard to  to the meaning 'odd.'  1  the fact that the phonetic series formed by qi ^ element kj> )  •*]  Another interesting phenomenon is  is distinct from that formed by the simple  . The former occurs only in Type B syllables, and the latter only in Type  A syllables. This could be seen as further supporting evidence for a different origin. 3  Dl  ^  is used in the present graph because both legs have to be depicted in order  to show that one is on one side and the other on the other side.  Usage It occurs only in the name Zi Qi  242  ^  "5" "SJ"*  'Prince Qi.'  JGWB 4676 (unidentified) 'some identify as gj j&> '  S.40.4xl  JWJ CunyiA542: gj %\  SW9a.6a: tl,  l l ^ .  &  \  , £  f  '-Though what we have in this graph is not exactly djl 'only having one leg.' When read ji, me also yj 2  .  4  &  , but  ^  ^  #1  JI  , which could depict the idea of  means'odd (number).' Professor Pulleyblank has pointed out to  'to lean' and, even more intriguing, fli 3% 'one-footed' (GSR lc').  Though there is some semantic overlap, as Professor Pulleyblank has pointed out to me, e.g. e.  'slope' and y i 'lean.' F o r an explanation of these terms, see Pulleyblank 1977-78:183-5. 3  ^  244  SWDZ 9a. 14a: Now w r i t t e n t % Tr: Qj  means chou  .  'ugly.' It consists of xi6  fl  'head' signific and QI  phonetic. Today in driving away pestilence 'ugly-heads' [i.e. masked shamans] are used.  Analysis This identification was made by Guo Moruo. He quotes ZhouliXia.guan.Sima:  "Inspecteur de region ou preservateur universel: II met une peau de jeune ours, orn6e de quatre yeux en m6tal jaune." (Biot 1851.n:225; Lin Yin 1974:324)  to which the Zheng Xuan commentary says: -fa  like the ugly-heads  of today.' As you can see, the graph shows a person wearing a mask. L i Xiaoding cautiously puts it in his Cunyi ^1 A  -0L  'doubtful' section, because the identification as Qi  is purely on inference. There is no solid evidence. However, one can still see that d l has to be used in order to portray the mask from the front.  Usage Name of a city.  2381 $  .  $  S.364.1xl68  JWJ 13.4039: hulng  JGWB 1606: huang fa  245  SW 13b.l7a: |> £  ^  7  &  ^  t t  t  , ^  ^  . &  ^  i§£  . £  ,  •  is the colour of the soil [i.e. yellowish-brown]. The character consists of  'field' and vfl . $ .  ffl  . Us ©  . . . . f ,  Tr: Huang lj[ tjjn  , }tb  is also phonetic.  >sL is the old form of guang  'light.'  Analysis In bronzes, this graph is often used to refer to an item of ceremonial clothing conferred upon people by their superiors. It was probably a belt containing rectangular plates of jade. The graph presumably represents a person wearing such a belt. Since the belt goes horizontally round the waist, it seems possible that the word h6ng 4fC 'horizontal' is related.  ^  is used because a belt going round the waist is most  conveniendy depicted from the front, rather than from the side as * \  or * ^  .  Usage Already in OBI, huang i s  is used as a rebus for 'yellow.' However, its main  usage is in the name of the ancient minister Huang Yin 1? sacrificed to just like the royal ancestors.  2422  ^  JGWB 1253: jigQ 3C  S.373.4x9  Not in JWJ, but cf. 10.3157:  =jiao  , who is regularly  246  SW10b.4a: Tr: Jiao  , £  J&  J &  . &  *  .  . JI  ...  means 'to cross the legs.' It is based on djL ;7v [representing a person],  and depicts the legs crossed.  Analysis The graph shows a person with the legs crossed. D | ^  is used because the front  view is the most convenient for depicting this. It would be very hard to depict this from the side.  Usage Obscure.  2423 ^  .^  JGWB 4689 (unidentified)  S.373.4x5  Not in JWJ  Analysis The graph consists of jiao  with a mouth-like element between the legs. A  friend of mine, John Lankford, suggested to me that it depicted the female pudendum.  247  Tsung-tung Chang (1970:54, n.7) also came to this conclusion. The 'sickness' context clearly shows that it refers to a part of the body, so this identification is certainly quite plausible. Although the sick person is not specified, it is a female ancestor that is called on to cure the affliction, so one may perhaps assume that the sick person was a woman, and this further supports the idea that it depicts a specifically female organ, though one should note that, although there is a tendency for sick people to be exorcized to ancestors of the same sex, this is by no means hard and fast. For example, of the 23 'exorcize Lady Hao' inscriptions at S. 140.2-3, 11 mention an ancestor/ancestress, of which 6 are male and 5 are female.  Chang identifies the present graph directiy with the modern character bj jfo , but this is something that is hard to prove. Professor Takashima has suggested to me that the graph depicts a womb, rather than the external pudendum. If one interpreted the element here as a frontal view of shen of the  a  , rather than jiao  , then the location  element would be inside the body, not between the legs, and this would make  Professor Takashima's suggestion preferable to Chang's. I  ^  However, note that yun  'pregnant,' which shows a baby inside the womb, is depicted from the side. I  think that in the present graph the  element is intended to represent an opening, and the  frontal view is used because the opening is at the front. Shima (1958:333) simply states that the present graph means a  HI.  'stomach,' but offers no supporting evidence for this  interpretation. From the context, the most one can say with any certainty is that it refers to a body part.  Usage In the two inscriptions from Yibian 4540, it clearly refers to a body part:  248  Renxu-day cracking, Dun tested:  Exorcize the sick body part (to)  Ancestress Gui.  Exorcize the sick body part to Ancestress Gui.  249  Section iii: Miscellaneous  250  154  $  JGWB 3: tiM ^  S.29.4xl (excluding  JWJ 1.13:  Man  SWla.la:  J  , |f  j&. £  |j  &  .  ^  -  , *  .  Tr: Tian 7s. means dian l | g 'top.' It is that than which there is nothing higher. It consists of yj — 'one' and d l  ^s 'big.'  Analysis According to Karlgren (GSR 361), the graph depicts an anthropomorphic deity. However, I can find no solid evidence that the ancient Chinese regarded Heaven as a person.  1  It was rather a very abstract concept, and certainly cannot be pinned down to  such a concrete notion as 'God.' None of the analyses recorded by L i Xiaoding even refer to anthropomorphic deities. I prefer Wang Guowei's analysis:  -Tv J2J^ ^  |J|][  k  originally referred to the human head, and that is why the graph depicts a person.' He further says that the head is emphasized because that is where the focus of the meaning is. One may compare such graphs as jian *% I |L 'to see' and wen  / f^j 'to hear,' where the eye and ear are disproportionately large to  convey the ideas of seeing and hearing. Chen Banghuai expresses a similar opinion, and quotes Zhang Binglin : 2  is a person's head, and by  Though some of the references to tian in the Shujing and Shijing, Professor Pulleyblank has pointed out to me, seem to be quite anthropomorphic in the sense of implying volitional activity. An 1  example from the Shijing would be "fc. ^ £  (HY 55/236/4) "Heaven looked down upon the  world below." Tian does not figure much in OBI, where the chief god is Shang Di P , but does figure prominently in Zhou literature. Thus it is hard to know what it actually meant to the Shang. No reference given. f  2  251  extension refers to the welkin' (i.e. because it is above a person's head). Note also the name of the monster Xingtian  in the Shanhaijing.Haiwaixijing:  Severed Head contested his spiritual powers with God. God cut his head off and buried it under Changyang Mountain.  1  I have cited the text from Yuan Ke 1985:191, who has emended  from an original -fit  on the basis of quotations of this passage in other works. He mentions that tian ^ written ^  in OBI and £  in bronzes, saying that o  ~ •  is  depicts the head, and that  'head' was probably the original meaning of tian %L (notice that in OBI, tian %~ is written £ and  •  , and •  is the OBI form of ding T  is the bronze form of ding "T  , while in bronzes it is written £ >  , though Professor Pulleyblank has pointed out to  me that ding ""]" tejrj < *tan is unlikely to be phonetic in tian. k.  fen < *t'a n, since in 2  phonetic series initial t'- has to be separated from the series t/t'/d and tg/tc'/d? when it is exclusive in a series, as it is in tian jfc. , and another problem that has to be taken into account is that the tijin 7 ^ , series shows final -m in such words as tian  fern'; on the  same grounds, X u Shen's paronomastic definition cannot be taken as evidence of an etymological relationship with dian  ten <  *t9ji).  Thus, Yuan concludes,  9f'l  k.  means 'cut the head.' This monster was so-called because God had cut its head off. There is also an interesting usage of tian 5^.  in the Yijing. Karlgren says it means 'to brand on  the forehead' (GSR 361), but traditional commentaries define it as kunxing /U T T J 'cutting off the hair as a punishment.' The passage runs:  ^Shanhaijing 2  Tongjian  112b. M y translation.  Professor Pulleyblank is currently working on the idea that this t - actually comes from *x -, as may be }  indicated by the early transcription of Hinduka as tianzhu , and the word x i a n if. 'Zoroastrianism,' but he has not yet published this theory. If correct, it would totally preclude any relationship of tijn 7^. with ding  or dian  .  252  "A man's hair and nose cut off."  1  L i Xiacxiing also includes ^  as a variant of tian  . Luo Zhenyu analyses this  form as shang _ h over d l 7 v , and explains: A Jl  JH(, $| 7 v  » 7v ^  ^  J^2-> 'Heaven is what a person bears, heaven is above man.' Although I am not clear on the role of d l ^  by comparing it with the graphs ^  and ^  in this graph, one can at least start off  . The second graph is yi \j  'city,' and  shows a subject kneeling under a city wall. The identification of the first graph is more problematical, but from its context (S.10.4) one can at least see that it is a person's name (his death is divined in Zhuihe 58 and Jimbun 455). I would analyse it as consisting of rgn signific, indicating 'man's name,' and ding  0  /T  phonetic, with the  phonetic element attached to the top of the 'person' element in such a way as to suggest the head. Then it is necessary to determine what meaning the graph ^>  is intended to  represent. Is it intended to represent 'heaven'? Or is it intended to represent 'head'? If the former, then maybe d l 7v  represents the 'greatness' of Heaven, or maybe one has to  accept the 'anthropomorphic' theory after all. If the latter, then one should note that body parts are usually indicated with rin ^ 'head' should not be written ^  , so there would be no reason why the graph for  . The role of d l ^  in the present graph is something  that I must leave open.  Usage Mainly in the expression tian yi shang  ^  , which I would translate as  'the head city Shang' (rather than 'celestial city'). Cf. Tsung-tung Chang (1970:236): i H Y 24/38/3. Translation from Wilhelm 1967:149. .  253  "Auch die Schreibung fiir die "Haupt"-stadt Shang die Bedeutung "Kopf' zuruckzufiihren." If ^ tian  £  p|j  =  fP, |g}  ...ist auf  (S.30.1) may be admitted as a variant of  , then Chang (ibid:236, ex.17.1) is able to come up with one example of it  meaning 'head':  ,  h  #  #1  •  (Yibian 9061)  Gengchen-day, the king (cracking): It will not sicken my head.  159  J G W B 1261: f n  ^  £  S.30.1x34  JWJ 10.3249: fu.  SW 10b.8a: ft ,  £  i"  A  & f\  , -T  . kk -K , — v"/ |< #  St •  .  ffl  vk >\  A. A - , *t 9 5t £ • ft ...  Tr: Fu ^_ means 'adult male.' It consists of d l ^v. with a line representing the pin. In the Zhou system there were eight inches in a foot and ten feet in a yard. A man grows eight feet tall, and that is why he is called a 'zhang-high fellow.'  Analysis The SW analysis is probably correct. As already mentioned, according to the ritual texts (e.g. Yili.Shiguanli), when a man became twenty he had his hair bound up and a cap pinned on to it. This symbolized the initiation into adulthood. One may compare the  1  The rubbing is not very clear, but it seems possible that there is a M.  Y  after 'king.'  254  graphs ^  , variant of nji.  j | / 4? 'woman,' and ^  , variant of mu  I  'mother,' which seem to show the hairpin that was a sign of womanhood. The d l fa element here perhaps indicates 'an adult,' a man who has reached the age where he can stand as a full-fledged member of society, shouldering all the responsibilities that he is expected to. Note also that there is a graph consisting of rin with a bar through the top: 3\  ^  . This is generally taken to be an abbreviation of ££.  (see S.333.4), and in this case the bar indicates that the head is being cut off.  Usage Seems to be the name of a place and a person.  169 j£  JGWB 513: msi  S.30.4x25  JWJ 4.1323: msi  SW4a.l7a:  $  ,*^.M£,/A*.#fc*-$£$fe!l4j&, .  S W D Z 4a.35b: Msi H. , shan -||" > y i $ £  and yjm. | ^  are all based on the same  concept [i.e. as regards the significance of the yang % 'sheep' element.] Tr: Mei  means gan  'delicious.' It consists of yang jf-  'sheep' and d l  'big.' Among the six kinds of domestic animal, the sheep is the chief supplier of meat. [The characters] mei J|_ Analysis  and shan *  'good' are based on the same concept.  255  X u Shen gives the meaning of mei J|_ as 'delicious,' rather than 'beautiful,' but I do not think this can be regarded as the basic meaning. I suspect that X u gives this definition purely in order to enable him to make a pun on shan 'cooked food.' In the classics, mei J ^ ,  'good'~shan  $H  is used mostly of appearance, i.e. 'beautiful.'  This must be its basic meaning, and to describe flavour as mei J ^ , is simply an extension of this meaning. How then does the present graph represent the concept 'beautiful'? Traditionally this character is regarded as a semantic compound of 'big' + 'sheep,' because big sheep are beautiful (or delicious?), as the X u Xuan commentary explains. Duan interprets the sheep as having a 'good' meaning in the characters mei f(t , shan  , yj.  and you | ^ .  1  The association between 'sheep' and 'goodness-beauty' is a litde hard to grasp, and I feel sure that these characters can all be explained in some other way. It would be too much of a digression to explore them all here, but I would just like to mention the relatively simple case of yi  nia < *nals, which is probably the primary form of 2d h  x i a < *rj al (?) h  'sacrifical animal' (see Shirakawa 1978:13). The 'sheep' element represents sacrificial animals in general, a n d w £ 4^ Qaa* < *nal'? is phonetic. Since wjj>  /  depicts a  weapon, it is possible that it is also intended to suggest the idea of cutting the animal up. Karlgren describes the present graph as "a man with a head adornment in the form of ram's horns" (GSR 568a). However, when we look at the bone graph, we see that it does not actually have yang ^ It occurs to me that usual OBI form is  /V>  / j^- 'sheep' on top, but some other unidentified element. tfi  could be an old pictograph for m6i /@  'eyebrow' (for which the  ). This would of course be phonetic: mei /g  and mei  differ  only in tone. On the other hand, if /V really is a headdress of some sort, then it could s  represent the idea of a person dressed up and looking fine.  ^ofessor Pulleyblank has suggested to me the interesting idea that the 'sheep' element may derive a 'good' significance from the word xiang  'propitious.'  256  U s a g e Name of a prince, Zi Mei ^ -Jsk  173  /Jj  . Apart from this, its usage is obscure.  JGWB 644: gji  S.32.3x62  JWJ 5.1725: QU  SW5a.20b:&  , A  *8 &  &.  /A *  , U $  Tr: O u -^T refers to people leaving each other.  . K/  ...  It consists of dJL k  signific  [representing a person] and rni L i phonetic.  Analysis Karlgren (GSR 642a) suggests that the graph may possibly represent the idea 'castrate' (i.e. qushi  ^  ), with the intercrural element representing the excised  portion. On the other hand, L i Xiaoding accepts the SW analysis. Although the element usually represents a mouth in OBI, there are cases where it has to be recognized as a container, e.g. fou. ^  / 1& 'earthenware vessel,' or zi tt* / ^  , also an earthenware  vessel. I feel more inclined to the opinion of Shang Chengzuo, who takes the whole graph as the primary form of mi -ifc , which is described in SW (5a.20b) as a food basket made of osier wicker. SW gives ^ that the  as the primary form of this character. This would mean  element here represents, not a person, but a lid. I think this is quite plausible,  since there are other graphs in which this element can only be interpreted as a lid, e.g. certain forms of M 'j£  (see JGWB 1254):  257  (Yibian 2924)  (Qianbian 5.5.5)  There is also a similar graph (JGWB 5603), which occurs only as a person's name but may perhaps be identified with the SW (10b.5a) character yun j[  (seal form: | |  ) the  original meaning of which is obscure, but which clearly depicts a similar kind of vessel:  (Kufang 1506 = Zongtu 20.2 ) 1  X u Shen places his M $ : j^L ^  ^  radical shortly after his d l 7v  'the d l ^  radical, and says: Jk  7v  element depicts the lid' (SW 10b.5a). I would say that the  present graph is the primary form of g l i  'lid.' One may explain the disproportionately  large size of the lid compared to the vessel by saying that this is the focus of the graph.  U s a g e It is possible that in some cases it means 'to leave,' but the context is too limited to be sure. It also occurs frequently in the phrase mi kyi $\ ^ intepreted as 'abandon lances.' Compare the phrase gj kui 2  / -jfe  , which could be  ^ \ (S.6.3), which could  be interpreted as 'raise lances.' Also a place name.  iThe same graph is found on the carved antler Kufang 1989. Tsung-tung Chang (1970:120, ex.7.24) translates this phrase as "Kui verlassen," taking kui as a place name here also. 2  258  177  JGWB 1300: M -,K  ^  S.34.1x25  JWJ 11.3359: M  SW lla.23b: $ A, Tr: Tai ^  >f|  .  Al 7^ , ^  $  .  means 'to wash rice.' It consists of s M  ^ - 'water' signific and d l "K.  phonetic.  Analysis The graph consists of d l  surrounded by drops of water. P I ^  is phonetic,  and at the same time suggests a person washing. Its primary meaning was probably 'wash,' and it later came to be used specifically for washing rice. The character is now generally written ' A , and survives only metaphorically, as in taotai impurities, purify.' The character t | i ^  'wash out  is really the same thing, but with two hands  added, as the seal form t ' H shows (SW lla.26b), thus suggesting the idea of washing something with the hands.  Usage Name of a person, Z i Tai %T  1074  'Prince Tai.'  &  S. 176.3x4 2428 A S.374.3x4  JGWB 1238 (both forms): chi  fa  259  JWJ 10.3197 (both forms): chi  ,  SW lOb.lb: #  |j&  ^  &  J  &  .M  , KL^JL.  JI  Tr: Chi ^ J N 'red' is the colour of the south. It consists of djt 7 \ 'fire.'.... ^  .... t  :  4 £.  'big' and huo  ^  : the old form consists of van y£_ 'flames'and i i . i - 'earth.'  Analysis The graph consists of 'big' and 'fire,' so the SW semantic compound analysis could well be correct. If so, then this is the only graph I have come across in which d l ^  actually stands for 'big,' i.e. 'big fire' = 'blazing red colour.'  Usage Graph 1074 occurs in a hunting inscription (Yibian 2908 = Bingbian 284.4),but it is hard to say what it means there. Graph 2428 on the other hand occurs before * i j ^ 'male horse' (Houbian 2.18.8), so it may be interpreted as referring to the colour of the horse. Shima's separation of these two graphs is perhaps meant to imply that the former contains shan iU  'mountain' while the latter contains huo  elements are often hard to distinguish in OBI.  3016  „ „  U ^  S.469.1x5  JWJ 10.3245: xi  . 4  JGWB 1260: xi  £  'fire.'  These two  260  SW10b.7b:|  , *  S W D Z 10b.l8b:  ^ X  SW here takes xi \  ,Mf l . S t H . in the sense of 2U. $$.  [SW 9b.l4b] which,  doubled, forms a descriptive epithet for the fat stomach of a pig. Tr: XI  means 'big belly.' It consists of dJL ^7v 'big' signific, and  phonetic.  abbreviated  is the zhduwen form of 2d ^ .  Analysis In his effort to account for the d l "A. element, X u Shen maintains that the original meaning of this character is 'big belly.' However, as I have already mentioned, there are hardly any bone graphs in which this element might reasonably be claimed to mean 'big.' Here, as usual, it simply represents a person. There is a rope attached to his head, and a hand holding the rope. Luo Zhenyu is thus probably right in claiming that the original meaning of 2d ^  is zulli  'convict,' or at least some similar meaning. Karlgren  (GSR 876d) gives the meanings 'slave, servant.' It is possible that the ^ ^  element is 2d  'to bind,' as X u Shen and Karlgren maintain. This would make it both signific and  phonetic, since the noun 2d %i may be interpreted as 'one who is bound, bondsman,' and may thus be related to the verb 2d ^  'to bind.' Note that the latter is in the departing  tone, which is often a sign of etymological derivation. actually A  1  The bone form of 2d ^  (JGWB 1546), but it is possible that the shorter form $  is  , which is the  ancestor of the modern form, already existed. It is possible that the graph ^  (S.469.1,  not in JGWB or JWJ) was the original form, and that the top part was then phonetized to 2  2d H I & . Y u Xingwu maintains that $  really depicts, not a rope, but a queue, and  One would normally expect the noun to be derived from the verb. However, as Downer (1959) has shown, the situation with qusheng derivation is now very confused (though presumably when it was still a piece of living morphology native speakers had a clear idea of its significance). 1  2  occurs in a Period I divination by diviner Que RsL (Jimbun 443), and  in a Period I  divination by diviner Zheng ^ (Zhiyi 18, S.468.4), whereas J^. occurs in Period I V and V inscriptions, so this could reflect a historical development in the graph.  261  that this hairstyle was affected by non^Chinese, and was thus a sign of being a barbarian. L i Xiaoding approves of this. However, the same element also occurs in ^ qiang  , a variant of  , and here it clearly depicts a rope tied round the neck. Whether or not the  barbarians wore queues in Shang times is certainly a moot point, and not the sort of thing for which archaeological evidence is ever likely to be forthcoming. Note that the present grapheme has a variant probably indicating 'type of human.' Perhaps the d l ^  element in the present graph  indicates that the slave was an adult male. However, there are no examples of this graph in its original meaning that could help us determine what sort of slave it referred to.  Usage Name of a hunting ground. There are no examples of it in its original meaning.  262  Part 2: ty  Ni ty I ^  is d i ^  upside down. It occurs in only one other graph, where it  is augmented by a foot and/or a road radical, and I think really this represents just another aspect of the same basic meaning, so there is not really much to say about this element. However, it is necessary to include it for the sake of completeness.  263  1341 V  JGWB 265: nl  $  S.212.4xl6  JWJ 3.687: ni ^  SW3a.2a: \  (this reference not in S)  Mfc ^  , ^  . &  f  T  ^  f  ^  Jk!t>  SWDZ 3a.2b: Later usually written £2L . [Duan also has kan U ^*  • 'pit' instead of che  'sprout, and explains that it has the same meaning as in xiong \ ^ , representing a 1  treacherous pit.] Tr: Hi  f  'going against' means 'not going with.' It consists of gan -p  SW ibid, as fjtn  [defined at  'to attack, aggress on'], underneath which a sprout opposes it.  Analysis The S W analysis is based on the corrupted seal form. As Luo Zhenyu notes, the bone graph is in fact simply djL ^  upside down. This is still clear in the bronze form  ( O / N J W B 252/315 ), though in the bronze form of n i i ^ 1  ( O / N J W B 184/217) it is  already quite corrupted:  However, Luo describes the bone graph as 'showing a person entering from the outside' ( TT- Ada  i&  $Y s\  ^ L - 7J7v  )> which seems unnecessary. I think rather that the  being upside down simply symbolizes the idea of 'contrary, opposing.' The  meaning 'to meet' that ni 15*.  sometimes has in the classics is simply another aspect of  this basic meaning, i.e. 'to meet head on.' 1  Though note that it only occurs on Shang vessels in very short inscriptions so there is no contextual  corroboration here that this really is m <^ •  264  One may compare the graph formed by turning r£n ^ *  7  /  upside down: dian  . This means literally 'turned on the head,' whereas ni "<f  does not have  this literal meaning.  Usage Obscure.  1340  $  JGWB 168: s i  &  S.212.4xll  JWJ 2.521: ni  ^  SW2b.3b:i5f  ,i&^.^^,##.8flfc©i£  a  >  181 ©  10, .  S W D Z 2b.5a: N i \ £ , and ying IkE. begin with the same sound [i.e. *rj-], and are interchangeable, e.g. Shu.Yugong " Tr: N i  means ying  signific and ni  1 , 1  is written "  " in the jinwen text.  'to meet.' It consists of chu6  [a walking radical]  phonetic. East of the Pass they say ni U £ . , west of the Pass they say  yingifll . N i I5i. rjiaj k < * n r a k and ying  Oiajn < * n r a r j  differed only in the nasality  of the final consonant. Xu shen regards this variation as dialectal.  Analysis  tongjian  6.853-4.  265  The bone graph consists of ni  on top of zhi  $  / j h . 'foot.'  While the  former is undoubtedly phonetic, the meanings 'oppose' and 'meet' are probably different aspects of the same basic meaning. In soft texts the character ni J3c_ is used in both meanings, but note that ying l£E  has only the meaning 'to meet' and does not have the  meaning 'to oppose.' The nasal~non-nasal variation in the final of related words is quite common in Chinese, though its significance has not yet been explored. One wonders if it is purely dialectal or also derivational. JGWB and JWJ both include the graphs $  and $  (S.324.4). The first form  is the direct ancestor of the modern character. Although the ni ty' element is phonetic in the present graph, it also seems likely that ni li£L is simply an augmented form of ni f  . Other members of the same word  family that Professor Pulleyblank has pointed out to me are against,' and  'to oppose,' I X  'to go  'go against, encounter,' all pronounced wji (GSR 60g, 60h, 58o).  Usage As L i Xiaoding notes, it sometimes means 'meet,' and is also used as a person's name and a place name.  266  S U M M A R Y O F CHAPTER U  To sum up then, although there is a residue of graphs for which I have been unable to arrive at a satisfying explanation, in the majority of cases the use of the d l ^  element  can be shown to be due to the need or desire felt by the designers of the script to portray both arms and/or legs or both sides of the body, or due to the desire to create a symmetrical graph (e.g. wji  267  Chapter I I I : ^  Part One: X  Of the three elements \  ,  $\  and ^  , only the  \  element is used  symbolically. The graph depicts a person kneeling, and is used to symbolize the ideas of yielding, inferiority and submission. This much seems fairly clear, but there is a great problem over the identity of this element. It evolves into the modern form "P , and SW has an entry for this. From X u Shen's paronomastic definitions of various characters containing y  , one can see that he thought it was capable of standing for ji£ t P , and  the Tangyun reading in SW accordingly gives it the same pronunciation as jig £p • However, there are no examples of  1  as an independent character, and it is highly  doubtful that Xu Shen knew its true pronunciation, since he manifestly did not know that it depicted a person kneeling. The mystery of the  ' P element is by no means solved, but I  offer various theories in my analysis thereof. After discussing the graph  \  itself, I shall divide the present part into three  sections, according to the role of this element as I have determined it, thus:  i. Reflects the posture that would be adopted ordinarily in real life for a given activity ii. Symbolizing yielding~inferiority~submission  lr  [nat is  $tz. "fcfl  . The Guangyun reading is the same.  268  iii.Phonetic (?)  269  256  JGWB 1089: jig"  f  S.41.1x32  JWJ 9.2865: ji£  f  SW9a.llb:  ,  ?  *p -# « 4 i f ,  #  Ai.  t  i  nm  *  ffl  £  f  ,  £  J® P  %  « » f , t it JH f -p, ^  SWDZ 9a.31b: The shape of the tally reflects what there is most of in that place. The information on tallies here is from Zhouli.Zhan.gjie [see Biot 1851.1:333-5]. Tr: Jis  means 'jade tally.' The person in charge of a country uses a jade tally, the  person in charge of a town uses a horn tally; the governer of a hilly district uses a tiger[shaped] tally, of an agrarian district uses a human[-shaped] tally, and of a marshy district uses a dragon[-shaped] tally; [keepers of] gates and passes use a bamboo tally, customs officers use a seal-tally, highway officers use a flag-tally. The character depicts [the act of] fitting together.  Analysis depicts a tally, i.e. it is the primary form of ji£ | ? P . However, the bone form clearly depicts a kneeling person, not a tally. Since 'f  1  is not  after all the primary form of ji£ j f p , this also calls into question whether it is pronounced likejM  • The identity of  J  J  requires investigation. I have been unable to find a hard  and fast solution, but I would like to offer the following three possibilities.  270  1.  ^  'to kneel.' This contains ji 2 J phonetic, and one  is me primary form of ji  could suggest that the sixth heavenly stem ji 2  /  is a stylized form of $  (for a  well-known example of the stylization of a graph that is commonly used as a rebus, cf. zhen H  / ^  < ding ^  kneeling, e.g. ?d under ji  / /fjf !). Bothji 2J  Tcnee'andji ^  and f  occur in characters to do with  'to kneel.' Karlgren classifies j i ^  k t ' < *ka'phonetic (GSR 953y), but if ji  function of the gi ^  gt  < *ga'  is phonetic, then what is the  gt < *gar element? Either ji 2J or gt  ^  would make a good  phonetic, but one hardly likes to say that here we have a character consisting of two phonetics and no signific. SW (14b. 10a) analyses the ji is based on X u Shen's understanding of j i 2 J hiding' (  L J element as signific, but this  as representing 'something curled up in Duan Yucai (SWDZ 14b.2b) explains  the SW definition of j i ^  , changju  with the legs spread out like a ji accounting for the cji ^  , i.e. to sit 'winnowing basket.' This would provide a way of  element as signific.  The problem with this proposal is that already in bones and bronzes j i Jt  is  written with ji C J :  (JGWB 1687 )  (O/NJWB 1851/2367)  2  If '2  were only the stylized form of  stem, then one would expect ji 3£  ^hat H  is from ^  ^  when used as a rebus for the sixth heavenly  to be written with ^  .  can be seen from (1) intermediary forms such as  (Jiabian 2304, JGWB  455: zhen A ), ^ (Xubian 5.16.4, JGWB 864: djmj rilr , but this example used as zhen i | ); (2) the fact that the full form is sometimes used (in fact, at S.396.1-4 most of the occurrences of ding TJ fv are used in the sense of zhen | | ^Example from Qianbian 3.18.4.  ).  271  2.  \  is the primary form of xi | i p 'knee.' This idea is put forward by Wen Shaofeng  and Yuan Tingdong (1983:317). They were inspired by a Qing philologist, Y u Chang *  t  I^P  (1854-1910), who claimed that the SW radical f  is the primary form of xi  (see S W G L 7.4019a-b). This suggestion has quite a lot in its favour. Firstly, xi  Jjjf  1  sit < *sac is phonetically very close to ji£  minking that  \  l %  ?P tset < *tssc.  If Xu Shen was right in  phonetically like ji^ f p , then this would support this identification.  Secondly, one could claim that also be signific) in jf ^  is phonetic, or at least a phonetic hint (since it may  / t p t s i k < * t s 9 c . This is all quite encouraging. It further  seems likely that xi l | t p Toiee' and j i £ ^P knee is a joint. In some inscriptions, \  'joint' are etymologically related, since the  is followed by a sacrificial victim, so it would  appear to be a verb here denoting some method of disposing of a victim. One could propose that it is used for the word gje. Against this idea is the fact that  ts'et < *ts'ac 'to cut.' ^  does not depict a person's knee, but a person  kneeUng. In the Shang script, body parts are usually indicated by emphasizing the part of the body in question, e.g. iiin ^  / -/fL ( %\ ) 'fundament,' and there is a graph ^  (see under I.l.iii.25) which seems to show the knee emphasized. There is no soft text evidence that xi  was ever used as a verb 'to kneel,' for which etymologically  unrelated words are used, such as j i the word xi  and gui  , though one could speculate that  was used as a verb by the Shang.  On the bone Jimbun 2283, the present element occurs twice, and in one occurrence it is written  , with a line drawn round the knee as if to focus on it. One may compare  zhou  'elbow' (JWJ 4.1507). Does this draw attention to the knee, or to the act  of kneeling? The context does not help us, as it is used here as a sacrificial verb, so it is probably a rebus for something.  272  3.  I  is me phonetic element m p i i I t , and pei  and f£i Pit.  According to SW, fei  both contain ji cL phonetic. This is clearly phonetically impossible.  The bone and bronze forms of p£i fOrL contain, not ji 2* Part). In this case then, it seems that ji  2J  , but "f  (see iii.2556 in this  is corrupted from " j . In OBI there is also a 1  graph %ft (see iii.880 in this Part), which I would like to identify as fei  . As for fei  , although the modern form is written with b_a HL> , the seal form (SW 4b. 15b) has 'P  . It is curious that these three characters should all be phonetically so close (they are  all in OC wei  rhyme and all have a labial initial), and all have the ^  common. Since the you  j § } 'wine vessel,' nix  'woman' and r6u  elements are clearly signific, it seems quite likely that the  \  element in 'meat'  element should be phonetic.  There are two other characters in SW that belong in this same phonetic group which X u Shen analyses as containing ji C J phonetic. These are p i - b i  b i ' (SW 13b. l i b ) ,  1  which one may hypothesize as coming from something like * b r a l ? , and fei~bi S J 2  p'uj ~ b i (SW lib. 15a, not in GSR).  I shall now discuss them.  Xu Shen defines this character as hui  x w i a ' < * x 9 l ? 'to ruin' (note the W  paronomastic intent), and this meaning is attested in Shu.Yaodian (Tongjian 1.324) in the phrase  (=  ) ^  ttj  , which Karlgren translates as "He neglects my orders,  he ruins his kin" (1950:3 f l l , see also Gloss 1238).  SW also gives the textually  unattested variant w-r , which Xu Shen claims contains pei though it seems rather obvious that it contains pli confirm the *-r- cluster, as pli simple element fei  4^  abbreviated phonetic,  phonetic (this variant helps to  p'e j < * p ' r s l then makes a better phonetic than the r  puj < * p e l would).  ^Foi the two modern readings, of which the first is irregular, see GSR 1237a'. The * - r - is reconstructed in order to account for the front vowel in EMC. See Pulleyblank 1984:26.  2  273  1 In the bronze script there is a graph consisting of Ui j £ . and "j ( 5Z> £  %  1  :  O/NJWB 1209/1503: 'not in S W ) . Since it only occurs as a person's  name, there is no contextual support for identifying this graph as bi i(L>. On the other hand, there is no bronze graph consisting of tji i-  and j i  competing for this  identification. These facts are too tenuous to constitute evidence, but they may perhaps serve a corroborative role when combined with superior evidence. Duan Yucai (SWDZ 13b.34a) says thatbi i-Zj is probably the same word as pi  )f!t  pi  < * p V e P , defined in SW (9b.5a) as beng m  p'aj -p'ut, defined in SW (ibid.) as bengsheng Rfl % the sound gloss'read like f£i~bi  \  'to collapse,' and pei~fu W> 'the sound of collapsing,' with  p'uj , b u j ~ p i < *p'ats, bets^prets' (this probably h  h  h  implies a pronunciation for Xu Shen's time of something like *p'9js). The j i in hi -*rL may be corrupted from I  £,  ^  element  .  SW puts this character under fei  signific and defines it as b j £ ^i]  'to  separate.' There are no textual examples of it in this meaning. The Guangyun gives the additional meaning 'an owl-like bird.' It occurs with this meaning, in the variant writing o  , in the ShanhaijingXishanjing, in the binome dufdi j# element in fei ffi-> is corrupted from  t£j  1  (compare how the ha  ).  Assuming that the SW definition is correct, this would mean that the j i  H-«  element would indeed have to be phonetic. This seems very strange, because the fei  ^  element would apparently make a much better phonetic than the j i other hand, if the j i S»  element. On the  element were signific, what area of meaning could it possibly  represent? The only meanings of ji  2J are 'oneself and 'sixth heavenly stem,' neither of  which seems to relate to the meaning 'separate,' (though these are of course both rebus meanings-it is not known what the graph depicts). Either as phonetic or signific, the ji  iSee Yuan Ke 1985:22. He says (ibid:42) that it is pronounced fii $ r l i .  274  ^  element in this character is mysterious, and suggests that it is in fact a corruption of  some other graphically similar element. Shen Jianshi (1960:451) notes that in one of the  &  _  Wang Renxu Qieyun manuscripts the character is written Q, > as if from yd  (a variant  writing of ' p ). In Long Yuchun (1968) it is under wei )%j rhyme written g^, , with the definition 'owl-like bird,' and given the alternative reading b i . h  If fei written # intoji  is signific in fgi , and  , then perhaps this word was originally simply  )> was later added as an elucidating phonetic, and became corrupted  CJ. The word ji I £ J  has both a different initial and a different final from the above  mentioned words (pli JUL , f£i  WJ , fe_i-£.(L , b_L ^"2J , fii  ), which between  themselves show a high level of phonetic similarity, so I feel very doubtful that ji C J is phonetic in any of them. For ggi IL while fj|i  there is proof that it originally contained }  still contains " P in the seal form. I feel inclined therefore to regard the  above characters as reflecting the true original pronunciation of \ in ji $h  ,  I «Jp  as signific rather than phonetic.  weighted against \  , and the role of  \  At least, the evidence seems to be  being phonetic in jf &f . Since a huiyi ^  interpretation of  this character is possible, I think that this interpretation should be adopted. If the phonetic element in the above characters really were j i 2J , one would expect more phonetic variety, e.g. a labial initial with a zhi  rhyme final, or a velar initial with a wei  rhyme final. Instead, in these characters that appear to contain j i L>  ^fi$L  phonetic, we find  two groups that are phonetically very consistent within themselves: a * P a l type and a *Ka-y type. This is probably the strongest argument for recognizing that the j i 2/  element in  these characters is a graphic collapse of two originally distinct elements. If this theory is correct, then what word does  represent? Its use as a radical  suggests that it means 'to kneel.' I have been unable to find a word with this sort of  275  pronunciation and this sort of meaning, but I hope at least to have shown that the identity of this element is problematical.  276  Section i: Real Life  277  265 j |  JGWB 1065: xiong %J  S.44.1xl22  JWJ 8.2801: xiong  %  SW la.4a: %%. , &  Tr: Z M %  £  f  ^  4  . ^  *  , &  k  , &  . -  © JU  means 'the person in charge of the sacrifice, the one who proffers the words  [i.e. says the prayers].' It consists of shi  'altar,' ren  'person,' and kou  'mouth.' One source says that it comes from djn %J abbreviated. The Yijing says: "Dui is mouth...is shaman."  1  Analysis JGWB and JWJ both include ^  as a variant of xiong ^  / /Li . However, I  think by now my exposition of graphic analysis will have led us to be surprised at, or at least to question, such a claim, and when we look into the matter we find that, while not completely clear-cut, there are cogent grounds for distinguishing these two graphs, the chief of which is context: their OBI contexts are demonstrably quite different. Sun Haibo himself (the author of JGWB) notes under the form jf| %  that 'xiong  is used as z h u % • ' Their pronunciation is so different that I hardly need give  reconstructions to demonstrate the impossibility of this. Other scholars have recognized I  a s a v a r i a n t o f z M f J / ^ (e.g. Long Yuchun 1968b:250), and this is certainly a  more realistic approach. I would say that $  shows a person kneeling in prayer (hence  the emphasis on the 'mouth' element), and that "1"$  is an augmented form with the  addition of an altar before the supplicant. I have found only three examples where definitely has to be interpreted as xiong !  H Y 5l/shuo/8  md52/shuoM.  \  , due to its being followed by a heavenly stem:  278  ^\  I % %  Houbian 1.7.10  $•  / £  Shiduo 1.423  $ $  I %  T  (ap. S.553.41)  Shiduo 2.294 (ap. S.44.4)  ZJ  If it is not considered too bold, I would suggest that these are 'spelling' mistakes. They are not the free variants that L i Xiaoding implies. Where then do all the J| from that are not marked by Sun Haibo as standing for zhji  in JGWB come  ? Upon checking, we find  that they are in fact all used as the name of a diviner (see S.569), so there is no evidence . We even find the graphs t  that these graphs stand for xiong  and 7  occurring  in the same inscription:  ••• I .»  J  ...  ^  I £  f t  1  #  •  (Jingjin40S0)  as far as Xiong Xin.  There are several inscriptions at S.44.1 reading ' |>  as far as such-and-such an ancestor,'  and this obviously parallels them. I have not translated the graph ^ is yet to be demonstrated. Further evidence for distinguishing ^  , since its meaning  from ^  , is the fact  that the graph TJJ[ is always written with a kneeling figure, never a standing one.  Usage As a sacrificial verb, H  appears to take an animal as its direct object and an  ancestor as its indirect object. The animal is often not mentioned, but the fact that it does sometimes occur shows that the ancestor must be the indirect object, even in those cases Unfortunately I do not have access to this collection. According to S, this inscription is the same as Ninghu 1.214, where however the graph for xiong is written in the normal fashion. Since this is a traced collection, it is not entirely reliable.  279  where it is not overtly marked with yj. "5"  'to.' If one interprets $  as zhii 4%, to  invoke,' then the animal would have to be re-interpreted as one step further down the hierarchy of obliquity from indirect object, i.e. 'invoke using such-and-such an animal.' This usage has been demonstrated by Chow Kwok-ching (1982) for what he calls 'Type A' verbs, e.g. gao  'announce' and Qjji  'seek' (the latter translated by Chow as  'invoke'). Some examples are:  r ^ ^ F  :  ^  £  T  +  •  (NanbeLMing 531)  We should make an announcement to Father Ting (with) one ox.  L ^ T i f f  •  (BucilAA)  On the next ting-hai day, we should invoke to Ting (with) two oxen. (Chow 1982:193, his translations)  Some of the $ the verb \  inscriptions are very similar to this. Although Chow does not discuss  , it does occur in one of his example sentences, where he transcribes it as zhu  , and translates it as if it were a Type A verb:  (Cuibian 148) If we announce the flood (lit. that the water entered...) to Shang Chia and pray to T'ai I (with) two oxen [sic: read 'one ox'], the king will receive assistance. (Chow 1982:232, his translation)  280  'Pray to' is also syntactically very similar to 'announce,' since in OBI grammar the direct object would be the thing prayed for or the matter announced. As Chow (1982:197) puts it, in the pattern 'Verb + Object Goal/Object Patient + (yi "f  ) + Object Beneficiary,' the  objects of Type A verbs are Object Goals (e.g. nian ^f" 'harvest'), while the objects of Type B verbs are Object Patients (i.e. sacrificial victims). If one interprets  \  as zhii  , then Chow's Type A concept enables one to interpret the syntactic roles of the various arguments that are dependent on it. However, none of the inscriptions at S.44.1 preclude a Type B analysis, since the direct object (i.e. the thing invoked or prayed for) is never mentioned. Thus one could still analyse  \  as a method of sacrifice taking the  sacrificial victim as its direct object. If  % and f \  represented the same word, then one would expect them to occur  in similar contexts. However, this does not seem to be the case. Whereas \  is usually  followed by objects (direct and indirect, or otherwise depending on one's interpretation), this is never the case with T $ verb yong  (see S.44.4-45.1).  is sometimes followed by the  'to use (in sacrifice),' and that is all; this could perhaps be interpreted as  'to sacrifice with prayers.' On the basis of the inscriptional context, one cannot do otherwise than say that  \  graph corresponding to zhu  and T $  have a different usage, f \  is clearly the bone  , so there is no reason not to interpret it as 'pray' or  'invoke' in the inscriptions, but it is possible that  % is used as a phonetic loan for some  other word. In the Guangyun, zhu  is included in a series of words that are all defined as  duan l f i / r 'to cut.' In his Guangya shuzheng, Wang Niansun quotes two passages from Chunqiu commentaries, " X_  ^  %% ^  " (HY 487/Ai 14/1 Gongyang) and " / ? £ | | v  " (HY 485/Ai 13/3 Guliang), and notes that the authors of these commentaries  both gloss zlnj. ffij as duan li?)p. As for the first passage, I feel that this is ambiguous, since zhu  here could easily be interpreted in its common usage as zhou 'to curse'  281  (now written  II  ), as one can see from the wider context, in which it paralleles sang  art-,  :4  Jt . *  When Yan Hui died, the Master said: "Alas! Heaven has bereft me!"  1  When Zi Lu died, the Master said: "Alas! Heaven has cursed me!"  The second passage however is more convincing:  Woo was a barbarian State, where they cut their hair short and tattooed their bodies.  2  If we intepret  j [ as meaning something like 'cut,' this would fit into the OBI context,  eg-  jj  ~~  4"  .  (Cuibian 339)  Cut up one ox.  I am unable to suggest what word it might be being used for.  283  f  JGWB 1112: gui  ^  This sentence is also found in Analects XI.8, so I have taken my translation from Waley 1971:154. The next sentence, about Zi Lu's death, is not found in Analects. translation from Legge 1861.5.1:81. 1  282  S.45.4x20  JWJ 9.2903: gui  SW 9,15a: t  , A  »  A  l . ^ A ,  &  Tr: Gui $L: that to which a person gui Jfp 'returns' is called gui  . It consists of r£n  'person' with the depiction of a spirit's head. The yin influence of spirits is harmful, [hence] the si  'secretive' element,  the old form contains shi *I» .  Analysis I originally assumed that ^  , somehow, depicted a spirit.  However, my  experience with graphic analysis has led me to think this over. There are many questions involved: How does it depict a spirit? Why is it kneeling? Why is the head represented by the ©  element? How does one explain the variant f ^  kneeling before an altar?  2  in which the spirit appears to be  A l l these questions make me feel unhappy with the usual  explanation that the graph depicts a spirit. M y own suggestion is that its use for &