UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Aspects of subordinative composite sentences in the period I oracle bone inscriptions Chow, Kwok-ching 1981-04-13

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

Item Metadata

Download

Media
831-UBC_1982_A1 C69.pdf [ 19.42MB ]
Metadata
JSON: 831-1.0095434.json
JSON-LD: 831-1.0095434-ld.json
RDF/XML (Pretty): 831-1.0095434-rdf.xml
RDF/JSON: 831-1.0095434-rdf.json
Turtle: 831-1.0095434-turtle.txt
N-Triples: 831-1.0095434-rdf-ntriples.txt
Original Record: 831-1.0095434-source.json
Full Text
831-1.0095434-fulltext.txt
Citation
831-1.0095434.ris

Full Text

ASPECTS OF SUBORDINATE COMPOSITE SENTENCES IN THE PERIOD I ORACLE BONE INSCRIPTIONS BY KWOK-CHING|CHOW A. The Chinese University of Hong Kong, 1972 Phil. The Chinese University of Hong Kong, 1974 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Asian Studies) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH -COLUMBIA November 1981 (c) KWOK-CHING CHOW In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Asian Studies The University of British Columbia 20 75 Wesbrook Place Vancouver, Canada V6T 1W5 Date 5 January, 1982 ABSTRACT Three types of subordinative composite sentence in the O.B.I., i.e., 'cause and effect', 'conditional' and 'simultaneous-successive', are investigated. Since there are no formal connective markers, the logical relationship between two clauses can only be determined on the basis of semantic considerations, the tui-chen pair, the practice of abbreviation and the larger context. A major type of 'cause and effect' sentence is the sentence of the pattern 'wu + V pu/fu ...' where the second clause represents some undesirable effect or situation. A consideration of the general positive versus negative pattern of the O.B.I, and the practice of abbreviation has led us to adopt the analysis 'cause and effect' for this sentence type. We may interpret sentences in which ch'i appears as conditional. Nevertheless, rather than a pure subordinate marker, the word ch'i is interpreted as a modal conveying the sense of uncertainty, a usage well illustrated in the classics. Also, the theory that treats ch'i as a marker of an embedded sentence has been refuted. The apodoses of 'conditional' and 'simultaneous-successive' sentences may represent an intended result or an undesirable consequence/situation. In most cases, these two types of apodoses can be easily distinguished. But in the case of raining, we have to rely on the idiomatic expressions 'y_u vii (wang yu) ' and 'kou yu (pu kou yii) ' in drawing the distinction. iii In divining about the appropriateness of a proposed activity, the 'conditional' and 'cause and effect' sentences both serve the same purpose. But an activity whose consequence is of greater gravity seems to motivate the employment of the latter. The ritual-sacrificial verbs can be roughly divided into two cate gories, type A and type B. Type A verbs represent major ritual-sacrificial activities requiring the accompaniment of ritual-sacrificial activities represented by type B verbs which can be placed either in front of or after the type A verbs. The latter case constitutes a conditional or simultaneous-successive sentence, while the former constitutes either a composite sentence incorporating a 'to clause' or a simultaneous-successive sentence. iv CONTENTS Acknowledgments Abstract List of Figures Symbols, Abbreviations and General Remarks CHAPTER ONE METHODOLOGY OF THE DECIPHERMENT OF O.B.I. PART ONE I. The Graphic Analysis II. The Limitation of Graphic Analysis and the Exploitation of Context III. The Use of Sources Other Than the Archaic Inscriptions IV. The Exploitation of 'Word Families' and its Limitation PART TWO I. A Review of Two Previous Works II. The Applicability and Limitation of Classical Chinese Syntax in the Grammatical Analysis of O.B.I. III. The Semantic Considerations IV. The Assumption of Sentence Parallelism and the Exploitation of Ch'eng-T'ao $ (Set) CHAPTER TWO THE IDENTIFICATION AND CLASSIFICATION OF COMPOSITE SENTENCES WITH PARTICULAR REFERENCE TO THE 'CAUSE AND EFFECT' TYPE I. The Criteria for Classifying Composite Sentences II. The Position of the Negative Wu tyj in Relation to the Analysis of Sentences V CHAPTER THREE CONDITIONAL AND SIMULTANEOUS-SUCCESSIVE SENTENCES (I) (NON-RITUAL-SACRIFICIAL VERBS) I. A General Description of Conditional and Simultaneous sentences II. The Contrast Between an Intended Effect and an Uncontrollable Contingency ( ^ $q ifa ) III. A Pseudo-Marker of the Conditional Clause Ch'i i% 1. The Sense of Uncertainty Conveyed by Ch'i }% 2. An Evaluation of the Theory that Ch'i jfe Functions as a Subordination Marker IV. A Hypothesis on the Rationale Underlying the choice between the Patterns 'V -77$' ' *rf] + V .. ., ftl 1 and 119 127 142 142 159 171 CHAPTER FOUR THE SUBORDINATIVE COMPOSITE SENTENCES (II) (RITUAL-SACRIFICIAL VERBS) I. A General Discussion of the Type A and Type B Verbs 1. The Contrast between Type A and Type B Verbs 2. The Relationship between the Type A and Type B Verbs 3. The Classification of Some Other Verbs 191 191 212 219 II. Analysis of the Structure '... V ... + ... V. III. The Simultaneous-Successive Sentence 226 248 CONCLUSION NOTES Chapter One Chapter Two Chapter Three Chapter Four 263 265 282 297 319 BIBLIOGRAPHY 359 vi LIST OF FIGURES FIGURE 1 PiUS. 20 87 a 2 ZlM 2287 b 3 Ch'ien $J 7.1.1 237 4 ^ I/J 5328 275 Ning. 'it 80 280 6 Tung Tso-pin 1933:173 324 7 I L 8810 330 8 Ts'ui 533 334 9 Ts'ui ^ 190 3310 Ts'ui 7^ 190 335 11 Chin JiL 711 338 12 K^u /j^ 714 9 13 Ch'i t X5 3314 1 340 15 18 1 vii ABBREVIATIONS, SYMBOLS AND SOME GENERAL REMARKS ^- Zero (non-existenc of an element) *X Reconstructed form or unattested unit X* The publication date of the first volume of a book X An unspecified unit A * B A may yield B A < B A may be derived from B A = B A is equal or equivalent to B A ^ B A is similar to B A vs B A versus B A B A versus B A / B A or B Q A single-graph lacuna in an inscription |2 A. lacuna of more than one graph tui-chen pair { G Graph W Word S Subject P Predicate S Signific P Phonetic The context will clarify the precise reference of the signs S and P. viii 0 Object OB Object-beneficiary OG Object-goal 01 Object-instrument OP Object-patient OV Object-victim V Verb N Noun VP Verb phrase NP Noun phrase R-S Ritual-sacrifice/ritual-sacrificial O.B.I. Oracle bone inscription ix AM  AYFC BMFEA BSOAS CHHP CKSYC  CKWT CKYKY CKYW CKYYY GYYY  CYYYChK CYYYCK EC HKCW HY HY JAOS KK KKHP KWTYC Asia Major An Yang Fa Chtleh Pao Kao "la . Published by CYYY Bulletin of Museum of Far Eastern Antiquities Bulletin of the School of< Oriental and African Studies Ching Hua HsUeh Pao ^J> CHWSLTs Chung Hua Wen Shih Lun Ts'ung $ ^ & Chung Kuo Shih Yen Chiu Nf7 (jjf) ^ff $L Chung Kuo Wen Tzu \^ k %• Chung Kuo K'o HsUeh YUan K'ao -Ku Yen Chiu So Chung Kuo Ytl Wen v|> 7\ X. Chung Kuo K'o Hslieh YUan YU Yen Yen Chiu So Chung Yang Yen Chiu YUan jfe #f ? J^L Chung Yang Yen Chiu YUan Li Shih YU Yen Yen Chiu So Chuan K'an  Chung Yang Yen Chiu YUan Li Shih YU Yen Yen Chiu So Chi K'an Early China Hsiang Kang Chung Wen Ta HsUeh Chung Kuo' Wen Hua Yen Chiu So  HsUeh Pao Harvard Yenching Index Harvard Yenching Institute Journal of the American Oriental^Society  K'ao Ku K'ao Ku HsUeh Pao -^jj Ku Wen Tzu Yen Chiu ^ ^-f MS Monumenta Serica Ping or Ping Pien See Chang Ping-ch'Uan 1957-72 San Tai SHKHCH  Sorui  SSCCS  SSTC SW/Shuo Wen TP WW YCHP YYYC San Tai Chi Chin Wen Ts'un She Hui K'o HsUeh Chan Hsien  Inkyo Bokuji Sorui Shih San Ching Chu Shu ^ t- ^"^ Shang Shu T'ung Chien ShuoWen Chieh Tzu Chu ^ %%{ 2r i& T'oung Pao Wen Wu fb Yen Ching Hstleh Pao "Sr *^ YU Yen Yen Chiu ^ \, "^?-f xi (.1) The correct understanding of an O.B.I, sentence structure relies heavily upon its tui-chen ^ counterpart and the large context (see Chapter One, p.69). Thus, the Pirig Pien (Chang Ping-ch'Uan ^ 7J^_ 1957*), a collection containing many of the fairly completely reconstructed plastrons has been selected as our corpus. Without speci fication, the word 'corpus' refers to the Ping Pien and statistics are based on the data collected from it. A great majority of the Ping Pien plastrons come from Period I; thus, the observations and generalizations presented in this thesis are virtually applicable to Period I O.B.I, only. Nevertheless, inscriptions from other periods are also adduced to serve the need of comparison or the need of illustrating certain problems. In addition, the Sorui has been used throughout the writing of this thesis. (2) Except those cited from other articles, inscriptions are normally provided with a word for word translation (each word separated by a slanted line /) and a free translation. In the free translation, the ch'ien-tz'u jfj (ff^ , e.g. ^ /Sf /<• , ptfL /=[ , is not translated since it is highly stereotyped and generally not problematic. In the original inscriptional sentences, the subject is normally unexpressed. Unless the context suggests otherwise, the author supplies in the free translation a subject 'we' which should be understood as a pronominal term referring to the Shang king and his closely associated people. Hstl-shu numerals are not reproduced. xii (3) Chou Fa-kao's ffy 1% fkj phonological reconstruction of archaic Chinese (in Chou's terminology, shang-ku-yin Y_ Ifr •% ) is employed in proposing readings for the archaic words(graphs). The author does not undertake to contend that Chou's system, reconstructed on the basis of the Shih Ching rhymes and hsieh-sheng-tzu ( ^ phonetic series of graphs), is completely applicable to the O.B.I, language. Obviously, there is a considerable temporal gap between these materials. Never theless, archaic Chinese represents the earliest stage whose phonology can be reconstructed with relative certitude. Thus, we have for the present no choice but to make use of it vigorously. When we say, for instance, that two words(graphs) have identical or similar pronunciations, we are, in effect, proposing that the presumed later equivalent forms of these two words(graphs) share an identical or similar pronunciation according to the reconstructed phonological system of archaic Chinese. Romanizations in front of a graph represent the reading in Wade-Gile romanization. Reconstructions according to Chou's system (marked by an asterisk*) came after the graph in question. The tone of a word is normally not indicated. In cases where there is the need to indicate the tone, the conventional signs qX (p'ing-sheng), °X (shang-sheng) and X° (ch'tt-sheng) are used. Such tone marks merely serve to indicate the tones of words in the ancient Chinese period (Chieh-y{ln period) . Whether this tonal character existed at all (and in what form did it exist) in the Shang language or archaic Chinese is an issue beyond the scope of the present thesis. xiii (A) The nomenclature 1 graph(word)/word(graph)' is used for practical expedience. As a general principle, the purpose of deciphering archaic inscriptions is to understand the language in which the inscriptions are written. In talking about language, of course, we have 'words' in mind. Nevertheless, at the first stage of decipherment, we are merely studying 'graphs' instead of words, although the ultimate aim is to discover what words these graphs stand for. In other words, we have first to study the strokes and other components of a graph in order to discover what later graph it may be equivalent to. Once the graph can be related to a later form, we have a clue to its phonetic shape and the graph acquires the status of a spoken word. From this point, we are, in fact, looking at it both as a graph and a word. The interlocked relationship between 'graphs' and 'words' is clearly exemplified in the study of graphic variants. The very name, 'graphic variants' refers exclusively to graphs. However, in order to determine whether two graphs are graphic variants or not, we must study the linguistic environment where they appear; to put it another way, we have to study them as meaningful words rather than as mere graphs (see Chapter One, Section II, p. 13). In other cases, we may feel justified in proposing a hypothetical 'meaning' for a graph through the analysis of its graphic structure while its pronunciation may well remain unrecoverable. For example, the graph i'^ has been tentatively interpreted as 'to decapitate (a captive)' but its prohunication is not clear to us . (see Chapter One, fn.8). Although •p^ definitely enters a sentence as a meaningful word rather than simply a graphic sign, it seems somewhat strange to discuss it as a 'word' xiv since no pronunciation can even be hypothesized. As a compromise, the author uses the term 'graph(word)1 to refer to the object in the process of decipherment. The use of this term is intended to show that the sign under study is being regarded simultaneously as a graph and as a word, in a process to discover its pronunciation by analysing its graphic structure. The term 'graph(word)/word(graph)' which may also be understood as 'word-character, logograph, tzu ' is here used in the sense of 'the word represented by this graph'. (5) As reflected in his translations of the O.B.I., an extensive and explicit interpretation of the various logical relationship between the clauses is first advanced by Paul L-M Serruys. The analyses such as 'conditional' proposed by Serruysai*e adopted throughout the present thesis without further acknowledgement. (6) For the sake of brevity, authorities whose work herein quoted are referred to simply by their names without any title.' XV ACKNOWLEDGEMENTS I could not have completed this thesis without assistance generously given by many individuals and institutions. I am principally indebted to Professor Takashima Ken-ichi, my research supervisor. From him I learned the general methodology and the appropriate approach to the study of O.B.I, grammar. Throughout the six-year period of inves tigation, refinement of the topic, research, drafting, and revision, his detailed and helpful suggestions were an essential impetus to progress. I was also greatly assisted by Professor E.G. Pulleyblank, who gave me much helpful advice on the phonological, morphological, and gramma tical aspects of this thesis. Professor Shimizu Shigeru of Kyoto University provided kind help during my period of research in Japan. Mr. Wong Yiu-kwan, Miss Yabuta Atsuko, Mrs. Winnie Louis, and Mrs. Vanessa Tam generously provided technical assistance in preparing the work for presentation. The Canadian Commonwealth Scholarship Committee and the Japanese Education Department (Mombusho) provided the financial support for my lengthy research. Last but not least, special thanks must be extended to Mr. Derek Herforth who listened, reacted, scrutinized draft after draft, and finally proofread every version from the first to the last. I deeply appreciate my fruitful discussions with him and his original comments; and offer deepest gratitude for his unfailing friendship and enthusiasm. CHOW, Kwok ching 5 January 1982 - 1 -CHAPTER ONE METHODOLOGY OF THE DECIPHERMENT OF O.B.I. Since the earliest studies of the O.B.I., Chinese scholars have developed their own research methods, methods which have never been made explicit. Lo Chen-y"u ^ ^ and Wang Kuo-wei . £ jJJJ ^£, to mention two of the earliest prominent O.B.I, scholars, did not establish a detailed methodology though they poineered in the decipherment of this archaic writing and undoubtedly laid a good foundation. In the preface of Ming  Yuan ^s, & (The Origin of Graphs), Sun Yi-jang|£ "ffL states, "Now I collect the bronze inscriptions, oracle bone : inscriptions, stone drum inscriptions and the Red Rock inscriptions found in Kuei Chou to compare and contrast with the ancient graphs in the Shuo Wen. (My purpose is) by illustrating their differences to show the origin from which they were simplified and changed as well as to group and classify (the graphs) in order to discover the general principles of the evolution of the ku-wen, 'large seal' and 'small seal' forms." a***^^ (1963: vol.1, p.2) From the above quotation, we learn that Sun's emphasis is on the historical study of the script. Undeniably, graphic comparison, taking Shuo Wen and other ancient philological writings as the point of reference (the basis for comparison) is the first step in the decipherment of the O.B.I., - 2 -but limiting one's study to graphic comparison will not always yield reliable results, since the meaning of a graph (word)-! must be determined' also by the context where it appears. Without taking into account the context, it is almost impossible to determine whether two similar graphs are merely free variants or whether they stand for two different words. Furthermore, interpretations of the O.B.I, involving phonetic-loan which, needless to say, can contribute to our understanding of the inscriptions are not hypotheses established by mere graphic analysis. These problems will be discussed in the following sections. The first significant discussion devoted to the methodology of O.B.I, decipherment must be attributed to T'ang Lan . In his book Ku Wen Tzu HsUeh Tao Lun ~& i~ % 1^9 » T'ang suggests several methods: 'the method of comparison ', 'the method of analogy $\ ' anc* ' componential analysis and its application *yf 'Z-'ty "^^T _jj§£^H J^J ' etc. Another important work concerning methodology is Lung YU-ch'un's Chung Kuo Wen Tzu HsUeh ^? |fSJ <L% ^ . The sections on the general characteristics of graphic structure, the evolution of graphs and graphic analysis are particularly helpful. It provides us with a basic knowledge of Chinese graphs and many concrete examples illus trating the methods we should employ in decipherment. While most of the Chinese scholars working oh the O.B.I, are actively deciphering inscriptions, they have not elaborated on the procedures of decipherment. The main task of their decipherment is to find a modern equivalent, if one survives, for the archaic graph in question. In most cases, when they propose that A (an archaic graph) is equivalent to A^ (a - 3 -modern graph), they are, in effect, proposing that A has a meaning, a pronunciation and a grammatical class similar to that of A^. Such a con clusion cannot be drawn, even tentatively, without going through the pro cedures explicitly put forth by Serruys in his article "Studies in the Language of the Shang Oracle Inscriptions': 1. What does the graphic concrete unit of lines or drawing actually represent? 2. What does this graphic representation stand for in terms of language, i.e., the word, which by syntactic position in the sentence may function as noun, verb, adjective or adverb... 3. What does this graph correspond to in successive periods of the Chinese writing; does it continue to appear in later texts, in basically unchanged form, leaving aside details of execution and graphic style, or is this graph replaced by a partially or completely new form (obtained by addition, transformation, combination of different elements, etc.)? 4. Finally, how was it pronounced or, at least, can we pronounce or read it according to its modern equivalent, once we have decided upon the various levels of identifi cation mentioned before? If we cannot propose any reading, then possibly, by approximation, we can find some sort of phonetic hint in the script, or some possible cognates in later characters of which the pronunciation is known. It should be noted that some questions remain even after the four questions have been satisfactorily answered, notably, the problem of phonetic-loans. either a verb or an adjective, corresponds graphically to the later form 7) and probably had a pronunciation something like *7"jwaY, the meaning of this word in the O.B.I, has not been satisfactorily elucidated. Only by comparing (1974: 19-20) Even if we know that the word-character expressions such as ty ( Z> $P ) with the phras next i-mao day' 32 0083 ) do we know - 4 -does not mean 'wing' but 'next'. (In other inscriptions J^? functions as an ritual-sacrificial verb with a still undetermined meaning.) The example shows that an analysis which reveals the original or pictographic meaning of a graph(word) may still leave undecided the question of that graph's meaning in a particular archaic inscription. The context of O.B.I, is the decisive factor in decipherment. Nevertheless, this does not mean that we can dispense with graphic analysis in our decipherments. A correct graphic analysis provides a base for an hypothesis of semantic ex tension and more importantly, a hypothesis of phonetic-loan. Had the graph (word) ^ not been identified as a pictograph of a wing ) which has a pronunciation something like *ri9k, the hypothesis that it is loaned for (*riek) would be difficult to substantiate. Following the procedures suggested by Serruys and making use of the methods proposed by T'ang Lan and Lung YU-ch'un, the author of the present thesis proposes the following general methodology,for the decipherment of archaic Chinese graphs. - 5 -PART ONE I. THE GRAPHIC ANALYSIS Before proceeding to details, there is a point that should be borne in mind, i.e., Shuo Wen has primary importance in graphic analysis. After the discovery of the O.B.I, and the recent advances in the study of bronze inscriptions, some scholars have come to believe that Shuo Wen can be discarded. Needless to say, the earlier a graph is, the more it reflects the original graphic structure (the meaning represented by the concrete lines) In this aspect, the definitions, explanations or analyses given in the Shuo  Wen have to be reevaluated and some of them must be rejected. Nevertheless, the 'small seal forms' ") * ^ , in comparison to the li^shu .ffii^* and k'ai-shu "Iffa -g* , are the forms derived most closely from the bronze inscriptions and the O.B.I. Shuo Wen serves as the stepping stone in proceeding from the k'ai-shu form to the bronze inscriptions and the O.B.I. Without the help of the Shuo Wen, many more difficulties, possibly insurmountable ones, would lie in the path of decipherment. Taking some simple graphs as examples: without the explanation and the 'small seal form' , there is no telling how much time and energy it would require for a scholar to determine that the O.B.I, form ^ represents a hand and has the reading identical or similar to the word(graph) ^ . To find out that the graph j (SW form ^ ) but not ^ represents the word 'man' would also take considerable time. And as a matter of fact, most of the archaic graphs which cause controversy do so due to the lack of correspondent forms in the Shuo Wen. There are other archaic inscriptions and writings that should also be taken as references,such as - 6 -the Stone Drum Inscriptions -fl |I$L and the bronze inscriptions both of which rely for their decipherment on comparison with the Shuo Wen. Of course, the analyses given in the Shuo Wen based on the 'small seal form' hundreds of years later than the O.B.T. and bronze inscriptions do not always accurately reflect the original graphic structure of the archaic graphs. Sun Yi-jang and Lo Chen-yU have pointed out many errors in Hstl Shen's er\ interpretations. For example, the graph (j^ is interpreted as "the bottom foundation, depicting the growing plants having foundation, so jj^ is taken as foot". But the O.B.I, and bronze inscriptions show that the top part of di is originally the depiction of a foot (Li Hsiao-ting, 1965:449). The ability to make use of the Shuo Wen without being misled by its errors depends on a careful study of archaic graphs themselves. As far as 'simple graphs J$%£J J^L ' are concerned, graphic analysis depends mainly on comparison with the 'small seal form' recorded in Shuo Wen. Thus JJ\ is 'grain', not 'water drop' or 'blood', because there is a 'small seal for.' |, % is 'rice-crop' not 'tree', because the Wll seal for„' has for 'rice-crop' and for 'tree'. As for the problem of 'combined graphs ^jf. 1^- ' , we have to employ the method proposed by T'ang Lan, i.e., the 'analysis of components 'TJ fttj ifftf % '. First we start by analysing a graph into different components, then we try to relate these components to later forms, i.e., the small seal forms. In cases where later forms for each component can be discovered, we may tentatively recombine these components in their later forms and try to look for a corresponding graph with identical components in the Shuo Wen. If such a graph can be found, the graphic analysis of an archaic graph can be said to have been completed. It goes - 7 -without saying that such immediate identification of an inscriptional graph with a Shuo Wen graph is not always possible. In cases where there is no similar form in the Shuo Wen, the hypo thetical meaning of a graph can only be deduced from the graphic structure. After a hypothesis is proposed, it has to be checked with the context (and sometimes with the classics). For example, there is a graph "^-^ which lacks any parallel in the Shuo Wen. This graph appears in the following sentence: king / prognosticate / say / have / harm / eight / day ^ 1*. 4^ tt% $*. (£|f-f keng-hsU /have / approach / cloud / from / east /.go through %+. f * % * yJ* it* ^ ifcrV ?f -• TJu / afternoon / also / have / go out / ^4^c( I from / north drink I at I river T'ung 426 The king prognosticated and said, "there will be harm". At the eighth day keng-hsU, there were approaching clouds from the east, they went through the Wu Nu star; in the afternoon, there was also an out going ^"^C from the north, (it) drank at the river.^ The shape of this graph may suggest many interpretations, it looks like a worm, a bridge, a jade ornament huang or a rainbow. From the context, we know that the clause 'there were approaching clouds from the east rfj 'occurs in front of the clause 'there was also an out going from the north jj}'>^&7^& *tf ' and this suggests that ' tfc ' may be a natural phenomenon. The inscription also reads 'drank at the river ' which suggests that tfC& is a kind of animal. In the classics, - 8 -we find the following passages: "Rainbow .... is also called ti-tung, it is often seen at the east when the sun is at the west; it is because it sips the moisture of the east." . j «* .... ,4U*tt a (Shih Ming Shu Cheng #f £i#c|&,p.l5) "Hung means ti-tung(rainbow), its shape is similar to that of a worm; (the graph) is derived from rj7 (worm) with X- as phonetic." (SW, p.680) "At that time, it was raining; a rainbow descended and, attaching itself to the palace, drank the water in the well." ^Kfa ,tfl>1%% (Han Shu, Vol.9, chdan 63, p.2757) "There are rainbows at its north, each rainbow has two heads." (Shan Hai Ching Chien Shu Jl &«$#L Vol.3, p.2) From the Shuo Wen, we know that a rainbow is considered to be worm like (the radical is ^ ); from the Shan Hai Ching, we know that the worm like rainbow is considered to have two heads; from the Han Shu and Shih Ming, we know that the ancients believed a rainbow drinks water. The shape of "$C$! looks like a rainbow with two heads; the clause 'drank at the river' also fits the description of $3. (rainbow) in the classics. Citing the above - 9 -evidence, scholars such as YU Hsing-wu ^4* (1940:15-19) proposed 2 that this graph should be interpreted as hung^J.. Such an identification seems to have answered the first two questions proposed by Serruys. That is, 1. The graph "^ty depicts a rainbow with two heads. 2. The word functions as a noun in the construction ' ^ & 'ficft ^ (there was also an out going rainbow from the north)'. > But can we say that this interpretation also answers the third and fourth questions? In other words, can we say: 3. The graph '^C^ corresponds to the later graph hung which was created by the hsieh-sheng method. 4. The graph "i^Cfi is pronounced in approximation as *gewng/*kewng/*krewng? Questions three and four are interrelated. Only after question four is answered can we say this graph corresponds to the later form As far as this graph is concerned, unfortunately, questions three and four cannot be answered conclusively. Even if we can ascertain that the graph signifies the concept 'rainbow', this does not necessarily mean that is pronounced as *gewng/*kewng/*krewng. There may be a synonym of the word hung ^feX-, a synonym which is phonetically totally unrelated to hung and yet represents the same concept 'rainbow'. In the Shuo Wen, Erh Ya and Shih Tsu 3f (T'-ung, jg^ ,". p.'. 86b-87), there are four different words, some mono syllabic and some bisyllabic, representing the concept rainbow: *YjwaK7*xjwaY *teSr-tewng % *keat-nj ier According to Kuo P'u If f%(ErhYa.pJB)^ ( ^ ff ) and ^are words representing female rainbows. The graphic structure of "^p"^ does not show the sex of the rainbow, so it is difficult to make a choice among the four words. As a matter of fact, even if the graphic structure of "7jff^ shows the sex, for instance, male, we still encounter the problem of making a choice among the words ^3- , ^ and 'j^ . In identifying the graph (word) TJC^ with the word ^X- , we are doing no more than finding a^" common word which represents the same concept as that of • Since the graph ^If"^ does not show a phonetic recognizable to us, there is no way to .tell whether the graph(word) was pronounced like^^ or like 1^ or like • With the limited material accessible to us, it seems we have to be content with the semantic identification of "tfcty with $JJ- , the most common word representing the same concept, and we should keep in mind that J^^i may be phonetically unrelated to . This sort of question is related to the practice of transcribing an archaic graph into a modern form. Traditionally there are 'six methods •yi* ' for creating a graph. Although scholars have divergent opinions con-cerning the original meaning of these 'six methods' and some even question whether the 'six methods' ought not to be reevaluated, it is generally accepted that a word may be represented by different graphs created by different methods. The concept 'spoon' can be represented by either ^ -11-Ca pictograph) or (a hsieh-sheng graph) both of which appear in Shuo Wen. In many cases, an archaic graph dies out and another graph comes into exist ence to represent the same word. For example, the word 'a pit; to bury (in a pit)' is written as ± ^(t/i) in the SW. In the O.B.I., there is a graph (Sorui 214.3) which depicts an ox buried in a pit and it is frequently used as a sacrificial verb representing the 'sacrifice of victims' to deities. Citing the sacrificial usage of the word k'an ^y? in the classics, e.g. Jjt ^% to bury the victim), Ch'iu Hsi-kuei ^? It proposes that the graph , where the component I J ( I I *kjam) appears to be the phonetic, is the primitive form of the word fil, *k'am (1980:162). (See also Yu Hsing-wu 1979:271.) Although the words L-J *kjam (t'an pu 'ft %f ) and *k'gm (ch'in pu ) belong to different archaic rhyming categories, they share a close phonological relationship. In addition to some cases of interrhyming among the words of these two categories, it is occasionally found that words from one single hsieh-sheng series appear in both categories, for example, *gjam/*giam -fife *kj ^ *griam *k/ (t'an pu #p *-am) (ch'in pu ?^,^P *-gm) And as Ch'iu (ibid.) has suggested, the word jVj? *griam(t'an pu) 'to trap, to fall into a pit' is obviously related to *k'am(ch'in pu) 'a pit, to bury (in a pit)' semantically. (Cf. the SW (p.695) uses the word *griam to define word jfe *k'am.) Notwithstanding the difference in the graphic structure of these two words(graphs), we may accept Ch'iu's theory that both t$ and A represent - 12 -the 'same' word in terms of the contextual and phonological similarities. In transcribing graphs such as , there are two traditions: one is mechanical transcription, that is, to analyse a graph into components and then to transcribe each of the components into its modern form. In this case, ^/ would become 5& • The other way may be termed 'semantic transcription', that is, to transcribe a graph into a modern form which represents a word assumed to be identical in meaning. In this way, ]^/ becomes jtfe . If we consider that the function of transcription is to tell readers what the 'meaning' of a graph(word) is, the 'semantic transcription' should have priority. But there is a serious drawback to 'semantic transcription'. Scholars often differ in opinion on what the meaning of a graph(word) is. There may be different modern forms suggested for transcribing an archaic graph. This not uncommonly produces a confusing picture. In the present thesis, however, since the original form is given in each case, the necessity of a mechanical transcription is not urgent and the 'semantic transcription' is employed. In some cases, both transcriptions are provided. - 13 -II. THE LIMITATION OF GRAPHIC ANALYSIS AND THE EXPLOITATION OF CONTEXT Graphic analysis is only the first step in the decipherment of O.B.I., for the validity of a graphic analysis must be verified by some other means. In the early states of O.B.I, studies, the graph was interpreted as $2-'to stand' since it graphically resembles the 'small seal form' ( JL. 'to stand'). From the Sung Dynasty until the late Ch'ing the graph ^ of the cyclical characters (kan-chih ^ ^ ) was a notorious conundrum because combinations such as kuei-tzu ^ ^ (San Tai /fa , chflan 9, p. 14, Ke Po  Kuei -$4: \fa f &.) and ting-tzu ^ (San Tai ^0 » chUan 9, p. 7, Shih  Sung Kuei 9^ ^fify which never occur in the traditional sixty kan-chih combinations are sometimes found. Only after studying the context where these graphs occur were scholars able to reject the traditional interpretation which relied solely.on graphic analysis and to reinterpret as (king) and ^ as f?j (the sixth ti-chih graph). Not only can a graphic interpretation be verified or refuted by means of context, but the problem of graphic variants can also be solved by reference to context. It is widely known that archaic graphs do not have stable forms. The graphic variation in archaic inscriptions often causes controversy in decipherment. One of the most frequently encountered graphs ^ ( pT to divine) is found in the following variant forms, e.g. , ^ and Graphic evolution through the course of time certainly has to be taken into consideration and sometimes variants may be attributed to the style of an individual engraver. Nevertheless, the graph J| is found in different shapes even in the same period. The case of |) does not cause much trouble since it highly frequently appears in very limited linguistic environments. - 14 -But in the case of graphically similar characters which appear in varied linguistic environments, the paleographer must take up the task of deter mining whether or not the variants all represent the same word. In other words, the relative weight to be placed on a single stroke or component is a problem which cannot be solved by graphic analysis alone. While graphic variants are very common in archaic inscriptions, we also know that two dis crete graphs representing two separate words may be distinguished by very minor graphic differences. Following are some examples: 1. The forms ^ , ^ , ^ and lj<} occur in the O.B.I. Do they represent the same word? If we interpret them as mere graphic variants of one word, then should we also view the form as one of the variants? The occurrence in the bronzes of an 'intermediate' form ^£ between ^ and makes the problem more complicated. It is after careful investigation of the context that we can determine that the first four forms should be taken as graphic variants while the form represents a different word. , , 3^ and ^ frequently appear in front of ^ ( -j^ mountain, mountain god) and ^ ( river, river god) and they can be preceded by the negative ^ ( fW should not). On the other hand, frequently occurs between personal names and is never preceded by 'J . Therefore, we can posit with relative con fidence that the first four forms represent the same word while represents another word. 2. In the Shuo Wen there is a graph ^ ( tree) which can be inter preted graphically as the later counterpart of either of the O.B.I, graphs)^ - 15 -or ^ . Since the graphic difference between X and !^ is very minor, one might be inclined to take these two as graphic variants. Moreover, the graph /& has another form , £ has another from % and ^} has another form . From the viewpoint of graphic analysis, the identity of ^ and ^ seems to be unquestionable. However, the linguistic environments in which these two graphs appear are completely different. The graph ^ appears exclusively in the kan-chih combinations, while the graph /fC never appears in kan-chih combinations. appears as a place name, a personal name and as a verb with an yet undetermined meaning. Therefore, we may conclude that and /j^ are different graphs(words). Now we must deal with the question raised by Lung Yu-ch'en: Why are and ^ interchangeable in ^ % , ^ and ^ %^ , though kept quite distinct when appearing independently? There is a passage in Lung's book which is worth quoting in full: "....a shape(structure) which appears as a component may change into another shape(structure). Therefore, cases where two graphs are kept clearly distinct when they appear independently, even though they are interchangeable when appearing as components, must be due to the fact that the way of engraving(a graph) was comparatively flexible such that one was misidentified for the other. (Such interchange-ability) should not be adduced as evidence to show that they are identical graphs. For example, the graphs n and q are identical (i.e. representing, the identical word), and ^| are identical. However, (we) cannot equate to -fcf- nor -"t^ to j . It is often the case that distinctions between two graphs are carefully maintained when these graphs are found independently while they are often confused or used interchangeably when appearing as components in other graphs. People of the Han Dynasty wrote (the word) as , but (the word) j^. is never written as people of the T'ang Dynasty wrote (the word) J|| as (but they never wrote the graph) to (represent the,word) § .... The graph ^ is different from 7) , the former is the 'word-character (to pray), the latter is fi_, (brother). But after adding - 16 -the component (Q.B.I, form: "J ), the graph ^ can be inscribed as || Since the bronze inscriptions extend over a long period of time and are spread over a wide area geographically, such cases are so numerous that they cannot be exhaustively listed. For example, in graphs such as , and , the shape Jc (the depiction of a leg and foot) are all . engraved as (-fcr*woman); may be engraved as , the shape vt being confused with *f . : . X is different from ^ , but %% may be engraved as .'^^ , may be engraved as ffi- , and ^ as j % ." (1968:248-250) In short, we must be extremely cautious in interpreting two inter changeable forms as representing the same word if such interchangeability exists only when they serve as components but not as independent graphs. The context must be taken into consideration. 3. Traditionally, most scholars, Yeh YU-sen 5* j^o > Kuo Mo-jo %\\%-%. and Li Hsiao-ting ^7$%. (1965:587-592) among them, consider and as graphs representing the same word. (Cf. the graphs | and both stand for the word wu 5f ; they are graphic variants. See T'ieh 76.2; Ch' ien "jjjHJ 1.9.3 and 3.9.1 for inscriptions showing the interchangeability of these two graphs.) The first person to seriously question this identification was HsU Chin-hsiung His arguments are based on the contexts in which ^ and / ^ appear. According to HsU, there are three usages of the graph (1) a sacrifice "f-r £ M. Tt i% «f> 1 ^ 4875 to / Ancestor Hsin / exorcise / sickness (It is) to Ancestor Hsin that we should exorcise the sickness, - 17 -(2) an official title /elk I A .v ..M T'ieh 183.4 i-mao1 / crack / Tui / / yU shih .... yU shih. Ku /€• 595 (3) a tribe name I / should not / summon / YU Fang I should not summon YU Fang. On the other hand, the graph 1^, is frequently associated with hunting and military activities, e.g., (4) wu-hsU / king / crack / test / hunt / Ch'iang / go / come have no, / disaster / king / prognosticate / say A t ft 2-35-1 auspicious / this / _y_U / capture / game / four In hunting at Ch'iang, there will not be disaster in going to and fro. The king prognosticated and said, "it is auspicious". In this charioting, we captured four game. (The translations of the word ^ / in the above four inscriptions are based on HsU's interpretations. See HsU 1963:4-7. On the meaning of the terms Jf^ ^ and Jpp ^ , see page 28-33 of this chapter.) Based on this evidence, HsU posits that and represent two different words; the - 18 -former is the earlier form of (to exorcise),(in the official title yU-shih %&P £ ) and ^Jf (in the trit>e name YU Fang^Jf-^f as found in the I Chou Shu ^ ). The latter graph /£, is the earlier form of $f which Shuo Wen defines as 'to drive a horse At. & H. •'. HsU' s inter pretation appears to be acceptable when the contexts in which they occur are examined superficially, however, a more careful study would show that and 1^, have the tendency, though not necessarily, to appear in different periods. A preliminary investigation based on Shima Kunio's jf^ ^jP |^ Inkyo Bokuji, Sorui f- ^jji ^ illustrates that predominates in early inscriptions while is more common in later periods. The graphic forms and personal names found by the side of are frequently ^ , JL. , ? . T , $ , , X . ^ . , 9 • ladies | f£, $ #£. ^ , princes ? faf , f , ? W , diviners ^ , ^ and ^ . As HsU noticed, the graph is frequently associated with sickness and disasters, subjects most common in divinations from the early periods. On the other hand, the graph frequently co-occurs with graphic forms such , and with common expressions such as ' ^ ti it] ( ^2 ^ 7^ have no disasters in going to and fro) which are commonly encountered in later period inscriptions. And as HsU pointed out, I £j is often associated t'ien-hunting (© ), a topic not frequently divined about in early period inscriptions. Moreover, the graph 1^, very frequently co-occurs with §i to form the expression tzu yU Jjpp (in this affair) which is one of the criteria for assigning an inscrip tion to the fifth period. This tendency towards complementary distribution (in terms of period) of these two graphs is ignored by HsU. In fact, there are some instances of 1^, which appear in environments - 19 -very similar to those of the graph for example: A. . zr U/ ^, A _, Ch'ien fl'l 1.35.5 i-ch'ou / crack / yu-sacrifice / exorcise / to Ancestral Mother Keng(?) / behead / twenty / aromatic-wine thirty We should perform an yu-sacrifice, (and then) perform an exorcism to Ancestral Mother Keng(?) (with) twenty beheaded victims and thirty goblets of aromatic wine. (For the interpretation of the graph(word) , see Chapter Four, footnotes no. 1 and 1'4.) Al' r~i -J . 4j , M - Ak Hou Shang %% ±- 27.1 ting-ch'ou / crack / Cheng / test / exorcise / to Ancestor Hsin / ten / penned-sheep We should perform an exorcism to Ancestor Hsin (with) ten penned-sheep . A2'M| 11 f * ft 1-2 test / chia-ch'en / yu-sacrifice / exorcise ^ $ H 0 ci.) f from / (Shang)Chia - 20 -On the day chia-ch'en, we should perform a yu-sacrifice, (and then) perform an exorcism (starting) from (Shang)Chia. u Ch'ien 6.13.1 kuei-ssu / exorcise / Prince She On a kuei-ssu day, we should exorcise Prince She. Bl'n . tiv K ^ Ts'un fir 2.224 ting... / crack / Ch'tteh / test / should not exorcise 'T" / Prince Huai We should hot exorcise Prince Huai. night / Fu / take charge of / affairs / receive .... At night when Fu takes charge of affairs, (we or he) will receive (assistance). C1 T'ieh  i-mao / crack / Tui I ... I take charge of / affair ft 183. < will take charge of affairs. t - 21 -Thus we can see that two of the three usages of the graph mentioned above are shared by I ^ . Now the picture becomes clear. The graph is mainly associated with exorcising illness, repelling other tribes and taking charge of affairs, all of which topics appear most frequently in early period inscriptions. The graph is mainly associated with the graph and occurs most frequently in later period inscriptions. In addition to this, in a much fewer number of cases, I shares the meaning of (mostly in the sense of 'exorcising', there are only two, out of twenty three sentences where and %^ co-occur). If the 'subject of divination' (, the seventh criterion in periodization) can be used as a criterion to periodize these inscriptions, the sentences in which is used in the sense of 'exorcising' might be classified as belonging to the early periods. Thus we may hypothesize that these two graphs are by and large graphic variants each representing two words: (1) to exorcise, to repel (both derived from the basic sense to go against) (2) to take charge of ^ predominates in the early periods while I ^, occurs in the early periods but does not come to predominate until later periods. The following is a figure describing the way in which the graphic and semantic changes of and l£j might have taken place: i. to exorcise, to repel -> (to exorcise, to repel) (to take charge of) • to take charge of - 22 -The term can be understood literally as 'in this taking charge of (affair) ->in this affair' which is derived from the meaning of %^ (Jl) (to take charge of affairs). The practice of using ^ ^ in the sense of 'to exorcise' became rare in later periods perhaps due to a change of religious concepts. The later kings did not believe in the effectiveness of exorcism as much as their predecessors, thus we may suggest they are more rational than religious. The fact that the sacrifices to one's ancestors become mere formalities also suggests this movement towards rationality. A new meaning 'to welcome' appears in the bronze inscriptions (in fact, there are some cases in the O.B.I, where the graph can also be interpreted as 'to welcome', but we lack a larger context to determine which interpretation to choose). The graph is sometimes engraved as The meaning 'to welcome' is also derived from the basic sense 'to go against, to go in the opposite direction'. Some Chinese words, such as ying and ni jj^ , which in modern Chinese have the meaning 'to welcome' also originally had the sense 'to go in the opposite direction (to meet so and so The definition 'to control a horse' which is given in the Shuo Wen for the graph ( '<{t^ ) is generally represented by.!^ ( ) in the bronzes. In the bronze inscriptions, 'to take charge of and 'to welcome' are the prevailing usages of. the graph "T^ and j£j . The problem of and 1^, illustrates that while using the context as a criterion in decipherment, there are other factors that must also be taken into consideration. - 23 -III. THE USE OF SOURCES OTHER THAN THE ARCHAIC INSCRIPTIONS The decipherment of the graphs $ rf\ > % Sj^ and is mainly determined by means of the O.B.I, context which may be termed 'internal evidence'. We may, on the other hand, obtain evidence from literary source, e.g. the classics, which can be used as external evidence. The following is an example showing how the context of words occurring in the classics can be made use of in decipherment: The bronze graph , which looks almost identical to the graph '(j? ( -j^ yellow) in the Shuo Wen, frequently appears in the terms representing things bestowed to the vassals from the emperor: San Tai 2_ <f^ , chtlan 9, P.20) ^, 'fv /|C f # <SunS Kuei '** A# f /( ^ ^ W ibid., p.38) f t H>i f Jfe (Fan Sheng Kuei -| £ 0£_ /h^'/hffr ^g. ^ ibid., p.37) tjr vJg. ^ ibid., p.37) If the graph(word) is taken literally, it would not make much sense. In the Li Chi , a passage of the yd tsao states, "On the first occasion, (the emperor) bestows an orange sacrificial knee cap and a black top gem of the girdle pendant; on the second occasion, (the emperor) bestows a red sacrificial knee - 24 -cap and a black top gem of the girdle pendant; on the third occasion, (the emperor) bestows.a red sacrificial knee cap and a green top gem of the girdle pendant". (SSCCS, chuan 30, p.7) The graphs (words) *pjwet and fjl *p'wat both represent the same thing 'knee cap'. Thus, in the classical expressions 3y ^j/ ^ ^ M> and the bronze phrases ^ ^ ^ / , the first three words are almost identical. Most scholars interpret the graph(word) ^3 *gwang as a phonetic-loan for *grang in the classics. Once this interpretation is established, we have a clue to interpret the term l& £ (K'ai Ting j|£ ^ . San Tai , chtian 4, p.33). The graph looks very much like the graph 7^ » hut a term such as /^v does not make sense. T'ang Lan notices that the graph tai (bone form: /t^ , small seal form: ) has an alternative form //\> in the Shuo Wen. The graph evolution is • If a similar graphic evolution has taken place in the case of ^ , we would expect to find a graph something like * which would probably, in turn, evolve into the shape /f\ ( ft_> *kang) as recorded in the Shuo Wen. Thus T'ang Lan proposes that should be transcribed as f]_j where f\j *kang is another phonetic-loan for wj *grang (1965:21A). As a matter of fact, the graphic identification proposed by T'ang Lan, though it sounds reasonable, lacks solid evidence since the form * is found nowhere in the inscriptions. But, T'ang Lan has pointed out, the strongest evidence to support his interpretation, is not the graphic structure of ^ but the comparison with the phrase $n $Jln the classics (ibid.). - 25 -The graph may serve as another example to illustrate the use of classical parallels in decipherment. This graph appears in the following clauses: (Mao Rung Ting h 'A jf^ ^4 San Tai ==./ft* , chUan 4, p.46) 1^ l^t^a^L CK'e Ting <t # D/ >f /f *P ibid., chUan 4, p.40) ^ th± 1= (S/ffl (Chin Rung An f % ' \t t «1 ibid, chUan 18, p. 13) a Arguing from context, Wu Ta-ch'eng ifc nas proposed that this graph should have the meaning 'to put into order' ;^ j^, and interprets it as the original form of the word nieh ^^jp- which HsU Shen defines as 'punish ment ^ -fc/ ' (1918: Vol.5, p.5, Mao Rung Ting). If graphic and semantic considerations are taken separately, Wu's interpretation is very appropriate. However, the graph rarely appears in the classics. According to the Shuo Wen's analysis, /j? functions as the phonetic, so its pronunciations should be something like *sj iat which may be related to the word ^ *ngia_t. In the classics, ^ often has the meaning 'punishment'. Whether and ^ both originated from the same word stem, and how the meaning of 'to put into order' and the meaning of 'punishment' are related to each other are questions Wu fails to answer. Liu Hsin-yUan /c," >/j|;(.1902: Vol.2, p.33) and Wang Ruo-wei • 0 (1968: Vol.6, p.261) take a different approach in inter preting this graph. They propose a meaning identical to Wu's but a different pronunciation, which is based on a comparison with the classics: - 26 -ffijL (Mao Rung Ting '£ $f ) tff X__ (SSTC 360289) "... and aid their sovereign." (Legge, p.479) 1£ I? ^. *P (K'o Ting fjfay 1^ X_ 1SL (SSTC 34Q171) "maintained and regulated the dynasty of Yin." (Legge, p.480) \% ft 1- H ' (Chin Rung An f" £ J ) 1^ 5C ® ill (Shih Ching HY 37/172/5) "May ye preserve and maintain your posterity." (Legge, p.273) Wang suggests the following theory to explain the relationship between these graphs: "... this (graph ) is the original form of the graphs i )C_ and 1 which appear in classical texts. In the shih ku jffi |£ (of Erh Ya §f ), we find, 'i X_ means to manage, to govern, to assist and to rear'. In the Shuo  Wen, 'i_ means to manage, to govern; it derives from pi (as the signific) and has i X. as the phonetic' In the YU shu J% <g (of Shang Shu # ), it says, ' there a capable man, to whom I can assign the correction of this calamity?' This shows that the graph i^ XL. appearing in classical texts is written i_ ^ in the ku wen ^ found in the wall. The graph 1^ ^ is probably a distortion of . Due to the similarities in structure, ^ became ^ at first. Since i^ -jf^ and pi are different in pronunciation, later people add i yL__ (to ) as the phonetic' - 27 -If vtfc m * f £ ^ £ £ Cl. #f H X . t; £ , # , % t, : , ;e t, A A. # ^ /f; 4 # ^ f ^ X.. £ *i * )C % M t * ft. ^ $ £ %^ »* tym. A A. # ^ ** # if If ft , H * *P K. «A ^7 (1968: Vol.6, p.261) In view of the context and the positions where the graphs ( ^$f' ) and appear, there is no doubt that they represent the same word. Without the comparison with the classics, even if we can propose the meaning 'to put into order' for the graph(word) as Wu does, we can hardly discover its correct pronunciation. And such an identification would be impossible if we did not disregard the graphic difference between It and X~ (it goes without saying that .X^ is graphically completely different from . Even the graph , which is written in the Shuo Wen, is not very similar to %^ ). Failure to identify )>j? with .X^ would make the correct understanding of the morpho logical (and semantic) relationship between the graphs , X^. and ^ difficult. The word %^ belongs to the archaic rhyme category ydeh jcj *-at and X^. belongs to chi *-ar; they both have the initial *ng- while differing slightly in the medial. The reconstruction of ^ is *ngiat andX-is *ngj ar. Based on the common morphological relationship between the so-+ jut called ch'd-sheng -Z-. nf and non-ch'U-sheng words, one may hypothesize that both *ngiat ^ and *ngiar X come from a single word stem (according to Pulleyblank (1973: 112-114)., some of the Middle Chinese ch'tl-sheng words come from words with an morphological marker *-s added.) *Ngiat »Jjt appears nominally and has the meaning 'punishment' while *ngiar appears verbally and has the meaning 'to put into order' which may be morphologically (and semantically) derived from the word 'punishment'. Thus, the graph(word) found in the bronzes is apparently polyphonic. It is pronounced *ngiar when having the meaning 'to put into order', but it probably also has the pronun-- 28 -V r ciation *ngiat since it can function as the phonetic of *ngiat ^ (Shuo Wen form ) and as the phonetic of *sjiat (Shuo Wen form Jf^ ). Unfor tunately, whether there is any morphological relationship between the initials *s- and *ng- is undeterminable in the case of ^ and ^ since ^ appears only as a surname or the name of a kind of grass. There is no evidence to support Wang Kuo-wei's hypothesis that is a mistake for (). In the bronzes, the graph is engraved as either ^ or ^ , both still easily distinguishable from or ( j?^ ) . The graph (^ ) recorded in the Shuo Wen may have been newly created by combining as phonetic with §^ , which frequently has the meaning 'king, to put into order, punishment', as the signific in order to disambiguate the polyphonic graph . When citing the classics as evidence in deciphering archaic inscriptions, there is one important point that must be borne in mind; that is, one must not over-emphasize superficial similarities between the classics and the archaic inscriptions. We may again take the decipherment of the word ( , ^ ) as an example: HsU Chin-hsiung %^ i^,^pt interprets the construction ;yji shih ^ ^ ( \ty tSL/^ ) as an official title (1963:7). This is not the only possible interpretation nor is it the most convincing one when applied to all cases. The words ^ are frequently preceded, directly or indirectly, by the verb hu 2p (to summon, to call upon) or ling (to order). For example, Cho 1 hsin-ssu / crack / test / order / many or masses / _yjl / shih - 29 -The construction 'y_U shih' is open to two interpretations in the above sentence: the first is, as HsU suggested, an official title. If we follow this suggestion, the translation would be 'to order the many yU-shih'. However, such an interpretation goes contrary to the preponderant syntactic patterns characterizing the verbs hu ^ and ling when an object is expressed. These two verbs are generally used in the following two patterns: (1) hu ^ and ling ^ as the first verbal element in a 'pivotal construction', i.e., the object of the verb hu *f or ling^ is the subject of the following verb, e.g., " M$ f * $ 4 f '4 test / summon / Lady Ching / millet / receive / harvest (If) we summon Lady Ching to plant millet, we will receive harvest. Chin <fc 645 king / should not / order / Pi / lead / masses / attack / Hu(?) t* Statelet Hou Shang %&Jl 16.10 The king should not order Pi to lead the masses to attack the Hu(?) Statelet. In some cases, the noun(object-subject) between the verbs is unexpressed, e.g., - 30 -ft * ^ £ ch'ien f'\ 3'29'7 test / summon / millet / receive / harvest (If) we summon (somebody) to plant millet, we will receive a (good) harvest. (2) Hu ^ and ling ^ followed by a clause, e.g., M J t*^A' ^ .-=5 4* ? * - * test / next / chia-hsU / River(god) / not / order / rain On the next chia-hsfl day, the River(god)will not order (it) to rain. I. ~L> 3121 In the O.B.I, language, it is not often that a personal object alone follows the verb hu ^ or ling ^ . In such rare cases the pattern is usually as follows: ' \% I ^Ml + personal name + ^ / /t~ ' (It is/is not so and so that we should call upon/order.) But we do not find any cases of * & / ^ ^ + k\\ %^ + I ^ ' (*It is/is not the yU-shih that we should call upon/sommon.) Though it cannot be adduced as an evidence to refute Hsu's analysis, the absence of the pattern ' * {*jb / ^ & + ^ + I ^ 1 cast certain doubt on the nominal interpretation of the construction • So we are obliged to test the validity of interpreting the sentence ' ^ ^ Hrf ^' as 'we should order the masses to take charge of affairs.' Now let us inves tigate the construction starting with the bronzes: - 31 -In the Ta Yd Ting 7^ (San Tai £. /ft' , chuan 4, p.42), we find the following sentence: 'In taking charge of affairs, (you) do not dare to indulge in wine.' In the Ch'i Hou Hu ffv ^ % (ibid., chUan 12, p.33): ^ 4?«r 'Thereby, I take charge of the emperor's affairs.' In the O.B.I., to interpret 'yU shih' as 'taking charge of affairs' makes better sense than taking it as an official title in many cases, e.g. CD H| 7 f „ $h A K %i \ $ test / call upon / (?) / to go into / take charge of/affair HsU#| 5.16.6 '(We should) summon ^£$L(?) to come into (the court or the capital) to take charge of affairs.' The verb 'j_u 1 nas two common usages in the O.B.I., namely, a. /\_ has the meaning 'to go into' as in sentences such as A ( 3- K the king goes into) and 5- A ( £. \ $1 the king goes into Shang). b. /v! has the meaning 'to bring in' as in sentences such as ^ A E ( 't ^ 5- Ch'Ueh brought in three). To interpret ' K. ^% as ' to summon (?) to go into y{l-shih' does not make much sense. To interpret it as 'to call upon C?) to bring in an official iyti-shih also sounds very odd, since the objects being brought in are generally understood as impersonal, - 32 -such as turtle plastrons or sacrificial victims. (2) t*. & $ | %4b Chih Hsu 19 in./ here / burn-sacrifice / Mountain(god)/ take charge of / affair \ We should perform a burn-sacrifice to the Mountain(god)! and take charge of affairs here. The only case in which the term yU-shih ^ ^ has to be interpreted as a nominal phrase is Chia ^ 1636: perhaps / call upon / north / yd-shih / protect We should perhaps call upon the north yU-shih to protect . .. Even in this case, the term yu-shih ^ would better be analysed as 'the one who takes charge of...' rather than.the same official title as recorded in the Chou Li (HY 5/9a). Ch'U Wan-li considers the term y!l-shih in this, inscription refer to 'officials in general %. ->1 3§ % $L (1961:216). Hsu also interprets the term ^ as a tribe name (see also Kuo Mo jo 1977:270), an interpretation based on the comparison with a passage in the I Chou Shu vSLlJ] 1g which states,^ - 33 -In his interpretation ,of the construction HsU has not taken into consideration the linguistic behaviour of the construction ^ *jf and the words co-occurt-ing-..with'; it. £-rl4hf 1 (. £ ^'f ^fttfi Wai ^ 30 chi-mao / crack / king / order / yU fang Since the verb ling -J7 frequently takes a noun clause as an object or appears as the first verbal element in a 'pivotal construction', the tran slation 'the king will order (so and so) to repel the tribe' is more appro priate than the translation 'the king will order YU Fang.' ^ T *\JL U iktrt* TVui ^ 887 ting-ch'ou / crack / y_U fang If the term yU fang is interpreted as a tribe name, the above sentence lacks a verb, a rather uncommon construction in the O.B.I. (To be sure, there are some abbreviated sentences where the verb is unexpressed, but most of these verbless sentences also abbreviate the ch'ien-tz'u JEf'J ffl^ > e.g. 'ting-ch'ou crack'. In cases where the ch'ien-tz'u is kept, the verb is rarely omitted.) The problems of the constructions yU shih and yU fang have illustrated that in making use of the classics, one should not rely on superficial similarities between the classics and the,,0.B.I. while neglecting the internal structure of the O.B.I. - 34 -IV. THE EXPLOITATION OF 'WORD FAMILIES' AND ITS LIMITATION In the early stages of the O..B.I. studies Chinese scholars had a tendency to put too much emphasis on graphic analysis. They singled out a graph and tried to discover what the individual graphic elements actually represent. Undeniably, this is the first step and an indispensable procedure in decipherment. Nevertheless, a single stroke can represent various objects. The stroke — can represent the surface of the earth as in graphs J[ ( fief) and ^- ( soil, earth); it can also represent a hairpin as in the graph ^ ( -7v_ man) or merely a line showing the abstract relationship between two objects as in graphs v> ( __t_ above) and /^ ( f\ below). The relationship between two components of a graph can either be hui-i ja, , e.g. (depicting an ox in a river, generally interpreted as ch'en y/jj ), or hsieh-sheng ^ e*8- ff"^ (where water/river (signifying flood) is signific and "f" ( ) is phonetic, generally interpreted as tsai ^ ). Simply looking at a graph may yield various graphically possible interpre tations which have either to be substantiated or disputed. In addition to making use of context, the study of 'word families' may corroborate an interpretation and aid the investigator in his attempt to grasp the original sense of a graph whose meaning has undergone subsequent change. The decipherment of the verbal meaning of LjJ t one of the most frequently encountered graphs in the O.B.I., relies heavily on the appli cation of the notion of word family. The graph(word) may occur in different environments: - 35 -1. after the negative wu fy) ( ^ ); in front of the preposition y(l •j' C -p ); frequently associated with ancestral names and animal names• 2. between numbers (after the first number, there may be a counter, e.g. I'A. ^ ); in this environment, it is comparable to the word ju_ X in classical texts. 3. in front of a noun; in this environment, it is comparable to the word _yu. in cases such as yu Chou (SSTC ^^QQ^^ and yu chung "1^ ^L (SSTC 10Q086) in classical texts. 4. between a verb and a noun; in this environment, it is inter changeable with the graph(word) X. ( ), e.g. '"^C versus ^ ^ . 5. in front of a nominal construction and appearing as the counterpart of the word wang HC- in tui-chen "^f jf) pairs, e.g. r*f*? versus "C Ai^i) ; in-this environment, it is comparable to the word ju_ 7^ (to have, there is) in classical texts. 6. in front of a nominalized verb, e.g. *Ji 4"J' » ^ J in this environment, it is comparable to constructions such as ^-j (Tso Chuan HY 45/fe18/1 & ) in the classics. Based on the usage defined in 2, 3, 4, 5 and 6, the graph(word) ^ should have a pronunciation identical or very similar to either yu *^E) or _yu ^ . Based on the evidence of 1, ^ is probably a 'sacrificial verb'. The interpretation of LjJ ^ as a sacrificial verb and assignment of a reading close to ju ~fj^ or _yu_ to it was an achievement of the early O.B.I, scholars. Nevertheless, they failed to suggest a more concrete meaning for the graph(word) LjJ ^ other than the vague 'sacrificial verb'. - 36 -Since the context where 4^i appears does not provide any clue to a more concrete meaning of the graph ^ » the only method is to study the word family of which ^ is a member. Unfortunately, never functions as a phonetic in other graphs. Owing to the,fact that ^ has a pronunciation very close to yvi and y_u X_ (to be more precise,. 4^ is interchangeable with y_u )^ in many contexts), we might tentatively group [il into the word family of which y_u and y_u are members. A. *°Yjw3Jf a conjunction 'and' (-— '*two things held together' *°Yjway to have, to exist, to possess 4-— '*to hold' *°yjw3^ friends j>W defines, 'those who share a common goal are friends; (the graphic structure) is two hands holding each other l&f fej $L, AJ\ ^ iB 'two people held/linked together' *yjw3y°' ffi^j partners <• 'two people held/linked together' *Tjway0^f/jj to protect -f '*to embrace(?)' This word may be related to & *ri£k in the sense of hu-i || ' to protect 4 to cover, to embrace a fledgeling with wings'. (Cf. Shih Ching (HY 62/245/3)" % ±_ % >7J(, % $|L 1|| ^- * was Place<i on tne cold ice, And a bird screened and supported him with its wings." (Legge, p.468); also Shih Ching (HY 62/244/8) " Ujs. || -f~ And secure comfort and support to his son." (Legge, p.463) *JjjwaJf° (§f 'to limit, to enclose; an area surrounded by walls' «f 'to hoid, to embrace' V - 37 -Cf. A parallel semantic-extension in the Indo-European language: "OHG garto 'garden, etc. ... all as orig. 'enclosure' fr. IE *gher- in Osc. heriiad 'capiat', Skt. hy- 'seize, hold', etc." (Buck 1949:490-1) B. *Yjw9y° ^ to pardon (to release) *jJjW£]T j|_ a vessel for pouring wine '*to release, to let go' *°xws/ gifts, sent-out valuables This tentative word family has been set up with words(graphs) selected from the hsieh-sheng series *Yjwajf. Certain other words (graphs) from this series have been excluded since they do not seem to share a common basic sense with the words(graphs) listed above; thus the present writer hesitates to treat them as members of the same word family. There are some other words which appear to belong to this word family but incor porate different phonetics, e.g. ^ *riak and possibly ^Jfiwsk (to limit, a circumscribed area *to hold, to embrace). The reason for selecting words(graphs) from one single hsieh-sheng series is that the members of a single series are 'convenient samples' which are conspicious, they share the same phonetic and this fact guarantees their phonological similarity. In any case, the nine words(graphs) listed above suffice for our present purpose of assigning a basic sense to the word(graph) . It is not difficult to observe that the word family of ju )L / % is apparently divided into two groups. We may even choose to treat them as two word families. The basic sense of group A is 'to hold, to embrace', a sense which can be easily derived from the meaning 'a right hand —to hold with a (right) hand'. The basic sense of group B is 'to release, to let go'. - 38 -Although these two senses are opposite to each other, this is not, after all, a rare phenomenon in Chinese semantics. The so-called fan-hstln fj/^%']\ antiphrastic words are well recognized. One may suspect that there is some morphological process involved. But this is unrecoverable in the present fragmentary state of our knowledge of Old Chinese phonology. As far back as we can reconstruct them, all the words, with the exception of *xwa/, have identical initials, final and medials and in each group, there are both the so-called ch'U-sheng and non-ch'U-sheng words. In any case, with these two basic senses in mind, we may tentatively assign a more specific meaning to the verb 4^ • There are two ways of doing it: (1) ^ as a member of the second group meaning 'to release'; so a sentence such as ^ /y. f 7^ ^can be interpreted as 'to release a captive to T'ai Chia', or more smoothly 'to offer a captive to T'ai Chia.' (2) ^ as a member of the first group meaning 'to possess'. Takashima notes that there is a switch in the semantic structure of the word in the construction ' Lfr' ^ /f^j (there is disaster)' and ' ^ jfa (to (cause) to have captives)' . The former one is a stative verb the negative of which is wang -£l ; while the latter one is a controllable, transitive, action-process verb of which the negative is wu ^ . CThe lo gical subject of ^fc* ^ /fyfj (to have disaster) is a beneficiary while the subject of ^ ^ ^ (to cause to have captives) is an agent; the beneficiary of is the surface object, e.g. T'ai Chia.^ - 39 -It is by the study of the word family of yu ^ii that we can propose a relatively more concrete meaning, in contrast to the vague interpretation 'yu-sacrifice', for the word yu Since its occurrence outnumbers other sacrificial verbs, it would seem natural to hypothesize that the verbal meaning of ^jj is simply 'to offer, to present', rather than a specific kind of sacrifice (Yao Hsiao-sui 1979a:382). In other words, the concept of Lfc/ may be very general and may have a broad extension which includes concepts such as 'yu-sacrifice', 'burn-sacrifice' as extensions. However, there is not much evidence to support this hypothesis (i.e. in the Ping Pien, there is no sentence, nor any set of sentences, which can be interpreted as 'in offering( ) to so and so, we will yu-sacrifice/burn-sacrifice...' or 'in performing a yu-sacrif ice/burn-sacrifice to so and so, we will offer...'). On the contrary, it appears that •+/ and other sacrificial verbs are of the same 'level', i.e., the extension of the concept of does not include concepts such as 'yu-sacrifice ''^ " or 'burn-sacrifice '%! '. For example, 1 r ft * % ^\ X) kuei-mao / crack / Ch'tteh / burn-sacrifice / River(god) one / ox / yu-offer / three / Ch'iang / dismember-sacrifice three / ox Ping 124 (2) we should burn-sacrifice one ox to the River(god)i yu-offer three Ch'iang-tribesmen and dismember-sacrifice three oxen. - 40 -kuei-mao / crack / Ch'ueh / yu-offer I to I River(god) - f I # « = £ three / Ch'iang / dismember-sacrifice / three / ox •* t - - V + burn-sacrifice / one / ox Ping 124 (1) We should yu-offer to the River(god) three Ch'iang-tribesmen, dismem ber-sacrif ice three oxen and burn-sacrifice one ox. The order of these three clauses is altered in this tui-chen pair, thus suggesting that they are in a co-ordinative relationship. And each verb, i.e., burn-sacrifice, ty dismember-sacrifice and ^} yu-offer, takes a particular kind or a particular number of sacrificial animals. It is not likely that ^ is a general term with the specific verbs <)j^ and ^J> as its 'realizations'. If one insists on interpreting Ljr' as a general term, there arises the question of why the number and method of sacrificing oxen have been made specific, i.e. ^ _ ^ , Cjp ^ , while the method of sacrificing the Ch'iang-tribesman remains inexplicit. Not only does LjJ appear to be on the same level as other sacrificial verbs in a sentence as illustrated above, on plastrons where we have a larger context, ^} often appears in a position 'parallel' to other sacrificial verbs. For example: i-mao / crack / Ch'ueh / test / come / i-hai yu-sacrifice / Hsia I / ten / behead / and / five - 41 -4*? u ty % dismember-sacrifice / ten / penned-sheep On the coming i-hai day, we should yu-sacrifice fifteen beheaded victims to Hsai I and dismember-sacrifice ten penned-sheep. Ping 197 (3) next / i-ssu / yu-offer / Ancestor I On the next i-ssu day, we should yu-offer to Ancestor I. Ping 197 (8) These two inscriptions appear on the same plastron. Each sentence has a different verb and a different indirect object(beneficiary). The two verbs 'to yu-sacrifice and <±J "to yu-offer' appear in parallel environ ments. And there is no other indication, either from the sentences or from the larger context, that is a verb representing a concept which includes the concepts of and ^ . Examples of this kind can be easily multiplied, e.g., Ping 207 and 360. For a further discussion, see Chapter Four, p.205-208. While the usage of 'word families' offers helps in proposing a meaning to the graph in question, there is one limitation. Making use of word families to decipher a graph is a procedure quite different from the study of word families per se. Normally, in the study of word families, the meaning of each graph is clear to the researcher. Even though there are graphs which have different phonetics, they can still be grouped in a word family on the basis of semantic similarity providing they have close pronunciations. On the other hand, in deciphering a graph we are hypothesizing about its meaning. - 42 -It is very dangerous to assign the graph in question to a word family with other graphs which merely have an identical or similar reading. In fact, even within a single hsieh-sheng series, we often find members of more than one word family. Moreover, there are two serious methodological difficulties in the study of word families per se. (1) Word families are set up on the basis of semantic and phonetic similarities, but there are no formal criteria to determine whether certain words are sufficiently similar to be grouped into a word family. Intuition appears to be the only basis for judgement, hence opinion varies from one scholar to another. In this situation, affirming that two words belong to a common word family depends largely on the imaginative ability of the person in question. The problem is that the more imaginative a person is, the larger word family he can set up. On the other hand, in the case where one person does not agree with assigning a word to a certain word family, the only criticism he can make is something like 'too forced' or 'far-fetched', also a subjective j udgement. (2) The study of morphological relationships between members of a word family is still in its primitive stage. Little has been done in applying the morphological processes such as those suggested by B. Karlgren (1933, 1957), Chou Tsu-mo (1947), Chou Fa-kao (1953), G.B. Downer (1959), E.G. Pulleyblank (1973) and Mei Tsu-lin (1980) to the investigation of how members of a word - 43 -family are related to each other. Needless to say, the irregularities in morphological processes, e.g. the lack of any consistent pattern relating tonal and morphological changes as evidenced by the fact that hao ^7" , when pronounced in the ch'U-sheng ^ is a verb 'to like' but an adjective 'good' in the shang-sheng .t- ^ while tu_ , when pronounced in the ch'U-sheng is a noun 'a measure' but a verb 'to measure' in the ju-sheng also poses a problem. Given two words the meanings, pronunciations and grammatical classes of which are known to a researcher, it is possible to posit a morphological process between them. However, with two words of which the pronunciations are clear to the researcher, the meaning and grammatical class of one of them cannot be confidently deduced from the meaning and grammatical class of the other by applying the rules of morphological process only. Moreover, in deciphering an archaic graph, the exact pronunciation of the graph in question, e.g. whether it has a voiced or unvoiced initial, is always unknown. In this case, even if we have a good command of the morphological processes, the meaning of the graph in question cannot be thus deduced. The lack of formal criteria makes the setting up of a word family highly subjective. For instance, there may be no serious objection to the view that the following graphs all share the basic sense 'crookedness'. ii) *kew/*kjew/*gjew a hook, to hook <&J *kew a basket trap made with crooked bamboo ^SJ *kjew a crooked back, a hunchback *kjew the foot curled up due to cold weather *kew the ends of the yoke which press on the sides of the animal1s neck 'PJIJ *kew a sickle But do the following two alslo belong to the same word family as the above graphs? ^ *kew the wizened face of age tfpj *gjew an old woman As an old woman often has a hunchback and the wrinkles on a wizened face are crooked, we may posit that they belong to the word family of the graphs in the group. However, *kew ^ and *gjew "jfc^ also both share the sense of 'old'. Can we hypothesize these two graphs form a separate word family? There are three other graphs which have 'PJ *kew/*kj ew/*gj ew as the phonetic: ffij *kjew a colt ^aj *kew a puppy *gj ew a kind of mouse smaller than the ordinary mouse Do these three graphs form a separate word family having the sense of 'smallness'? Or should we accept Todo Akiyasu's ^jjj ^ 3^ /[^opinion that the sense of 'smallness' is derived from the sense of 'crookedness'? (1965:329) Supposing there is a semantically unknown graph having £j *kew/*kjew/ *gjew as a phonetic, what sense should we assign to1 it? In addition to the - 45 -difficulties associated with grouping or separating the 'intermediate words', i.e., words with no clear semantic relationship to others, of a certain word family, there is another problem, namely, that of 'graphs created through phonetic-loan i|[ ^ 9rf" Z. ^K^% ''"-6 The 'graphs created through phonetic-loan' may be a pitfall when making use of 'word families' to decipher a graph. The phonetic itself, when used in its original meaning, may have a 'sense' completely unrelated to the graphs in which it serves as a phonetic. For example, the graph(word) |^] *ts'rewng is the depiction of a window. Nevertheless, the hsieh-sheng series written with is composed of two word families which seem to lack any semantic relationship: A. . *tsewng to bring together /ffi,*tsewng the middle of (the east and west) steps (The basic sense is 'various things being combined together'.) B. ,r,/g>*ts'ewng green silk %*ts' ewng a precious stone of a greenish blue colour J^j£l*ts'ewng a black (green) and white horse (The basic sense is 'green'.) Suppose there is a graph having |J) *ts'rewng as the phonetic and the meaning of this graph is unknown; we then have two (or even three if YF) is also included) word families to which this graph(word) might be assigned. The semantic complexity of a hsieh-sheng series from which one or more root senses can be inferred causes great difficulties in exploiting the concept of word family to assign a meaning to an unknown graph(word). - 46 -It is a principle that all hypothetical meanings of words suggested by means of graphic analyses or comparisons with the classics must be sub stantiated by the context. This is particularly true in cases where hypothetical meanings are suggested by the study of word families. The interpretation of the graph (word) ( "f" > ) serves as an good example. The word family j( ( ^- , 7j£. ) has two basic and opposing senses: A. J| *ts'a/*ts'ra/*ts'ia/*sia/*ts'reY diverge rt- a o o— o o discrepancy *dza/*tsja disease o a— — '/^ *^sa/*tsa.° crippl•jfij^ *§zra cut tree*—*to make :s'a/*dza uneven teeth a 6 something incomplete *tsa/*tsa° left (Cf. sinister) (The basic sense of this group is 'discrepancy, unsatisfactory and to eliminate'.) B$Lm$.*tsa° to assist \ 3l *ts'a/*ts'ra/*ts'ia/*sia/*ts'reV to recover from sickness O & o o " (The basic sense of this group is 'to improve'.) (All these words are selected from Karlgren's Grammata Serica (1940: no.5) and Todo's Kan.1l Gogen Jiten } J ^ 1J ^ ^ (1965: no.153).) - 47 -Parallel to the word family, the word 7j. ( ^ ) is also used in two opposite senses in the classics: A. ^ ftz.-,ft£A-CTso Chuan HY 272/-j[l0/ ftp) Whom the son of Heaven favors, my ruler also favors, whom he disapproves, my ruler also disapproves." (Legge, p.449) /f> jfr. ~~pL ^ (Tso Chuan HY 365/9^4/ . \ 6£ ) 'Is it not improper...' (Legge, p.599) W, ~h A. -9 -t (Chan Kuo Ts' e: weits'e /fe ^ WW^W^' chuan 23, p.474) u '(Chang I) will, for sure, approach Ch'in and desert Wei.' B- f * & m i H (Shih Ching HY 82/304/7) '...who gave his assistance to the king Shang.' (Legge, p.643) (Tso Chuan HY 122/f-f23A ^2) ...your lordship laid your charge on Ch'ung-erh as to how he should assist the son of Heaven.' (Leege, p.187) It is on the two opposite senses of the graph (word) j( ( y£ 5 4^) that scholarly opinion has diverged. Ch'en Meng-chia (1956:569), Chang Ping-ch'tian (1957: vol.1, p.72), Serruys (1974:56), Keightley (1978:66) and Ch'en Wei-chan (1980:191) interpret ^ as -vi (to assist). Ch'U Wan-li (1961:304), Li Hsiao-ting (1965:951), Mickel (.1976:226), Takashima (in - 48 -a private consultation dated 1978), Li Hstteh-ch'in ^ ^ ^ (and Wang Ytt-hsin) ^. ^ fy^ (.1980:252) consider that ^ has a sense of 'violate, obstruct and pervert'. (In the Sorui, Shima does not give a separate entry to this graph. All the inscriptions where j( appears are listed under ^ ( j{ ) which, in some cases, is equivalent to 'to assist'.) Since word families and the classics do not offer help in determining the interpretation of the graph ^ , we have to rely on the context in which this graph appears: is *t §•% <u ^* y% } * t * king / behead / many / T'un / not / approve / obstruct / from lower / upper Ping 523 (1) 8 9 If the king beheads many T'un-tribesmen, there will be disapproval and obstruction from the lower and upper spirits. king / behead / many / T un / not / obstruct / approve / from lower / upper Ping 523 (2) If the king beheads many T'un-tribesmen, there will not be obstruction but approval from the lower and upper spirits. If the graph(word) ~f" is interpreted as 'to assist', the con structions ^ •j' and 7*^7^, maY be analysed as verb compounds (coordinative - 49 -verbs) and be translated as 'to approve and assist', 'to assist and approve' respectively. The question arising from such an analysis is why the order of these words is ~j~ in the first inscription but ~f~ ^ in the second inscription. Does it merely represent a stylistic variation? Though such a possibility, of course, exists, it is difficult to verify. It should be pointed out that these two inscriptions are engraved in a perfectly symmetrical position which suggests that they are tui-chen pairs. To interpret % 'j' and j~ -fa as verb compounds (coordinative verbs) would yield the following two translations: 'If the king beheads many T'un-tribesmen, it will neither be approved nor assisted by the lower and upper spirits.' 'If the king beheads many T'un-tribesmen, it will neither be assisted nor approved by the lower and upper spirits.' The positive/negative polarity expressed by most tui-chen pairs is absent here a very rare case in the O.B.I. On the contrary, if ~f~ is interpreted as 'to obstruct, to pervert' and a syntactic break is put between and y1" (and, ~f~ and ^ ), the positive/negative sense of a tui-chen pair is retained as shown in the translations provided on page 48. The variation in the order of the two words -fa 7 an(i / "56 is thus semantically required and not merely stylistic. (Ping 523 (1) and (.2) are the inscriptions cited by Mickel (.1976:226) who contends that the graph (word) ^ conveys a bad sense.) To interpret as 'to obstruct, to pervert' can also better explain why the graph(word) f~ is omitted from the second inscription of the following pair: - 50 -jen-tzu / crack / Cheng / test / we / perhaps / make / settlement *** A* fr ft ^U7(1) ti-god / not / obstruct / approve If we perhaps make a settlement, ti-god will not obstruct but approve (or, will be pleased). X*/*i n ^ «s *>. kuei-ch'ou / crack / Cheng / test / should not / make X* settlement / god / approve Ping 147 (2) If we do not make a settlement, ti-god will approve (or, will be pleased). (The 'conditional' usage of the negative wu ^) is proposed by Serruys (1974:56) and further clarified by Takashima (1977:53). The translation of Ping 147 (2) is in accordance with Takashima's interpretation (ibid.-54).) If f~ is interpreted as 'to assist', the positive counterpart of ' %p$p "f" (tne 8°d will neither assist nor approve)' should be something like '* rjjj -j- ^.(fthe god will assist and approve). But the inscription reads ' 7^ (god will approve)'. Stylistic practice seems to be the only explanation. Whereas if we interpret as 'to obstruct, to pervert', the ission of in the clause ^ is easily understandable, since '* 7^ (the god will obstruct and approve)' does not make any omiss sense. PART TWO THE APPROACH OF GRAMMATICAL ANALYSIS I. A REVIEW OF TWO PREVIOUS WORKS The study of O.B.I, has been going on for almost eighty years. While much concern has been lavished on the graphic decipherment of the inscriptions and the social structure of the Shang period, the grammatical aspect of O.B.I, language has not been carefully investigated. Hu Kuang-wei ^^in his book Chia Ku Wen Li >f ^ £ fa\ published in 1928, was the first to touch upon the grammatical problems of O.B.I, language. However, this book which contains many graphic misinterpretations is a very superficial study, the major portion of which does not deal with the O.B.I, grammar. 1. Kuan Hsieh-ch'u's ^ $71 Yin HsU Chia Ku K'o Tz'u Ti YU Fa Yen Chiu M f f £'J # ft tfj % In 1953, Kuan Hsieh-ch'u published the above book which we may speak of as the first work focusing specifically on the grammar of O.B.I, language. It goes without saying that at the time when Kuan wrote this book, the decipherment of graphs(words) which, is a necessary condition of grammati cal analysis was not as advanced as at the present time. Misinterpretations of graphs(words)} and thus misinterpretations of sentence structure, were unavoidable. Kuan should not be held fully responsible for the incorrect 52 -analyses of grammatical structures which are based on incorrect decipherment. Nevertheless, Kuan is not without fault for the serious weakpoints in his approach, that is, he based his grammatical analyses on certain unsound hypotheses for which he gave no documentation or proof and, furthermore, he ignored the implications of his analyses for classical Chinese syntax in general. For example, Kuan cites- the following inscription to show that the subject can be placed after the verb: it ft Y \> \£± M( tif? 0(JD f chi-ssu / crack / Ch'u / test / exorcise / king I to I (Shang)Chai -— Ts'ui $p 100 twelve month (Kuan, 1953:16) (Kuan's theory is adopted by Ch'en Wei-chan (1980:188).) If we follow Kuan's interpretation, the translation of the above sentence would be, '...will the king exorcise to Shang Chia...' with an inverted subject 'king ^J, '. In providing such an analysis, Kuan gives no reason for not interpreting the word 'king' as the direct object of the verb 'exorcise'. Bearing in mind that word order has particular importance in Chinese syntax, to interpret the word 'king' as the subject would require very strong evidence. The following question must be answered before one can accept the analysis that the word 'king' is the subject in the : inscription-5*' *ff 5> f X- f ' : In the O.B.I., there are numerous inscriptions having the pattern: 'exorcise + personal name + (to) + ancestral name', for instance, exorcise / Lady Hao I to I Father I (We should) exorcise Lady Hao to Father I. - 53 -In the pattern 'exorcise + personal name + (to) + ancestral name, ' is it always the case that the 'personal name' functions as the subject of the-verb 'exorcise'? (1) If the answer is affirmative, then a problem arises, namely, when the personal name is 'Lady so and so', there is no case where this personal name precedes the verb 'exorcise' (cf. the patterns ' jf, ...' and '2f-p £ ...* both can be found in the O.B.I.); in other words, the pattern is always 'exorcise + Lady so and so + (to) + ancestral name' as in Ch ien ^jj 1.38.2. Why must the order of the subject and verb be inverted in the case where the subject and verb are 'Lady so and so' and 'exorcise' respectively? (2) If the answer is negative, then one may ask what criteria are used in determining which personal names occurring in the pattern 'exorcise + personal name + (to) + ancestral name' are subjects and which (presumably 'lady' ') are not. Is there a semantic restriction that the word 'king' cannot be the object of the verb 'exorcise'? Unfortunately, we cannot even speculate on how Kuan would answer these questions since his rationale in analysing this inscription has not been made explicit. xs: Another inscription Kuan cites as an example of inverted word order —"f 1300 this / go out / Ch'iang / have not / misfortune (Kuan, 1953:16) - 54 -Kuan proposes that the word 'Ch'iang-tribesmen' is the subject of the verb 'go out' (the presumed translation would be 'today the Ch'iang-tribesmen go out, will there be misfortune (to us)). Kuan does not explain why the syntactic break cannot be placed between 'go out' and 'Ch'iang-tribesmen', i.e., 'in today's going but, will the Ch'iang-tribesmen not have misfortune', nor does Kuan explain why the word 'go out' cannot be interpreted as adjectival, i.e., 'will today's out-going Ch'iang-tribesmen not have mis fortune'. It may be because Kuan does not believe the Shang people would divine about the fortunes of Ch'iang-tribesmen (their enemies) that he has not adopted either of the alternate interpretations suggested here. But Kuan's consideration does not seem to have much merit. Although the Ch'iang tribe was an enemy of the Shang people for a long period, this does not exclude the possibility that their well-being was of concern to the Shang. There are indications that many Ch'iang-tribesmen were members of the Shang labour force. For example, t«r Up Hi ff It hs in-mao 7 crack / vg (?) / test / call upon / many / Ch'iang-tribesman / chase / (JI) I capture HsU ifif 4.29.2 (If) we call upon many Ch'iang-tribesmen to chase (?)» we will capture them. ..*?*• n &k t* If ft« eng-hsU / crack / test / I / order / (?) / follow ?fc fBffl fit ftm Ch'iang-tribesman / hunt / have no / misfortune I_ Z-> 4692 - 55 -(If) I order ^ (?) to follow the Ch' iang-tribesmen to hunt, there will be no misfortune. Although we cannot determine whether these Ch'iang-tribesmen were enslaved (Yang Hsiang-k'uei 1962:21; Yao Hsiao-sui 1979a:382-384) , or whether they formed a semi-independent group of Ch'iang friendly to the Shang, we may reasonably hypothesize that any misfortune which befell these tribesmen was considered disadvantageous by the Shang people. Thus, divining about the fortunes of the Ch'iang-tribesmen (enslaved ones) may be very similar to divining about the harvest. In any case, Kuan should not posit a solution which drastically violates the common syntactic pattern of the O.B.I, language and classical Chinese without giving reasons. Kuan commits another obvious error when he neglects certain peculiarities of the O.B.I, and proposes a theory which is extremely alien to the Chinese language throughout its development. Kuan states, "in the inscriptions, there are some animal nouns which have markers of gender. For example, i-wei / crack / exorcise / to / Ancestral Mother Hsin i «t ** ^**•,#* Tih017j 5328 Ancestral Mother Keui / hui / male pig / hui / sheep keng-yin / crack / ascend-sacrifice / chi-sacrifice /\ X, 1 Z, 5321 female pig / Ancestral Father Keng In the inscriptions, the (sign) _L found by the side of an animal' noun is the gender marker of male, ^ is the gender marker of female. , ^JL and are bull, male sheep and boar respectively; ^ , ^ and ^ are femaLjp ox, female sheep and sow respectively. In a pair of tui-chen sentences, there are also some (graphs) abbreviated to }) and "]" as found in example no.8: 'Is it a - 56 -female? Is it a male? Will we yu-sacrifice?' Mr. Kuo Mo-jo states, "What are _L and I) ? J- and h are the abbreviated forms of 'ancestor' and 'ancestral mother'... is in fact, the pictograph of phallus, so it can be abbreviated into _L ; tl> is the (semantic) extension of the graph (word) 'spoon' ^(jp Probably the female sex organ looks like a spoon, so £ is used (to represent) 'ancestral mother' and 'female'." It is still questionable whether graphs such as and fy^C* are monosyllabic. By analogy with graphs such as §\ ( ) 'Ancestral Father I' and 1(\ (Jttt % ) 'Ancestral Mother Hsin', tyi, and tyL, may be ho-t'i-tzu •" * **" T (1953:30) Kuan has not clearly defined the term ho-t'i-tzu. It is generally understood that ho-t'i-tzu are those graphs which are formed from two independent graphs and engraved as one single graph. They differ from hsieh-sheng graphs in that the former cannot be analyzed as the combination of two component graphs, one functioning as signific and the other as phonetic. They differ from hui-i graphs in that, in later traditional sources, each of their componential graphs is represented by an independent graph, e.g. the ho-t'i- tzu £J ( 2 + j ) is written as Pao I %jj[ ZJ in the Shih Chi |£, . Whether these graphs are monosyllabic or not is not clear.'''''' If graphs such as ( y + _L ) and ^ ( ^ + 4 ) are bisyllabic, i.e., is pronounced ^dziaV-^gjwa/, Sj^ is pronounced *bj ier/*bj ien-*ngjwa/, thenjL ( dr )*dzia^ and I) ( ^ ) *bjier/*bj ien are independent graphs (words). They should not be termed 'gender markers'. The primary criterion in distin guishing a 'marker' from a word is that a marker cannot appear by itself, i.e., it must be a bound morpheme. But Kuan has cited an inscription in which both 1 ( -t ) and j ( \^ ) appear by themselves: chia-shen / test / Hsiao I / chi-sacrifice / have no / harrir" hui / female / hui / male / should not / yu-offer / this / employ 12 Ning *jp 88 (Kuan, 1953:31) It appears that Kuan has confused the distinction between grammatical category and lexical category. The Chinese language, as far it is traceable, expresses sexual polarity lexically but not grammatcially. 2. Ch'en Meng-chia's ^%^^\ Yin HsU Pu Tz'u Tsung Shu « £ h %\ tltii As a specialist in the O.B.I., Ch'en avoids many of the errors involved in interpreting individual graphs. Nevertheless, his grammatical analysis of the O.B.I, leaves something to be desired. The most serious drawback is that Ch'en confuses semantic analysis with syntactic analysis, thus leaving alternate interpretations untested. For example, he analyses the inscription below in the following way: a9 al-fc «i fi- ^s: Dj sun / have / eclipse / perhaps / announce / to / Father Ting Ts'ui "4^ 55 (Ch'en, 1956:99) O1(S-V-O)-V-P-O2 P preposition 0^ direct object 0^ indirect object V verb - 58 -Although it is undeniable, from the context, that 'the sun was eclipsed' is the content of the announcement (the fact being announced), to analyse this inscription as one single sentence instead of two sentences or two clauses requires more evidence. Ch'en has provided none. It appears that he treats the inscription Ts'ui 55 in the same way as the following one: IIS TltflDf^T M.0 2.1..3 three hundred / Ch'iang / use / to / Ting (Ch'en, 1956:101) (It is) three hundred Ch'iang-tribesmen (that we should) use (sacrifice) to Tirtg. It goes without saying that constructions such as 'three hundred Ch'iang-tribesmen' can rarely appear as an independent sentence while 'the sun was eclipsed' can. Moreover, since the common environment where one finds an antiposited object does not appear, e.g. the initially placed particle hui jij^ , there is no obvious reason why 'the sun was eclipsed' has to be taken as the object. Ch'en has not considered other possibilities such as: 'Cause and effect': 'Since the sun was eclipsed, we should announce (this fact) to Father Ting.' 'Cumulative': 'The sun having been eclipsed, we will announce (this fact) to Father Ting.' 'Conditional': 'If the sun is eclipsed, we will announce (this fact) to Father Ting.' - 59 -In Ch'en's book, grammar takes up one out of twenty chapters. It is probably for this reason that Ch'en can only describe in outline what he believes to be the grammatical structure of the O.B.I, without going into detailed discussions. II. THE APPLICABILITY AND LIMITATION OF CLASSICAL CHINESE SYNTAX IN THE GRAMMATICAL ANALYSIS OF O.B.I. The formulation of a grammar depends upon the understanding of the relationships between words, which in turn depends upon the correct decipherment of individual graphs(words). However, as has been shown in previous sections, the decipherment of individual graphs relies heavily on the recognition of their linguistic behaviour, which in turn depends upon a correct understanding of the O.B.I, syntax. Theoretically, this constitutes a circular method of research. In other words, it appears that whether we are deciphering a graph or postulating a grammar of the O.B.I., we are forced to make assumptions about the other. Unsound though this method may appear to be, O.B.I, scholars have been using it in their study and have made significant accomplishments. Rather than starting from a solid, proven basis, these scholars work from certain assumptions, namely, that the O.B.I, is in the same lineage as the bronze inscriptions and Shuo Wen form graphs, and that the O.B.I, syntax is basically similar to classical Chinese. These assumptions have never been, nor are they ever likely to be, seriously - 60 -challenged. Since the end of last century, in the course of research on the O.B.I., no evidence has been uncovered which would lead to the discarding of these assumptions, though slight modifications and rectifications have been made from time to time. Thus, assuming that the O.B.I, syntax is basically similar to classical Chinese, e.g. that the general word order is SVO; and modifiers precede the modified, an inscription such as is interpreted as 'we will receive a millet harvest'. Of course, in the process of decipherment, one cannot expect a language several hundred years earlier than classical Chinese to be com pletely identical to classical Chinese (see Keightley 1978:66). Even within the scope of classical Chinese, there are pecularities among different texts. In the O.B.I., a highly restricted language dealing mainly with divination, it would not be unreasonable to expect to encounter some particular features not shared by classical Chinese. For example, starting from the Lun YU %^ and Tso Chuan /j: f^- , sentence final particles such as yeh and i ^ become almost indispensible in the classical Chinese system. Nevertheless, we do not encounter such particles in the O.B.I, (neither do we encounter them in large numbers in the Shih Ching %^ bf>£_, Shang Shu fcj and I Ching With the knowledge that sentence final particles do not, as a linguis tic feature, exist in the O.B.I., to interpret hu *f in the inscription 1 ft' 4( A iff ^ (*Should we sacrifice to Hsien Wu and HsUeh Wu?) as an 14 interrogative particle is very questionable. As Serruys points out, such an interpretation fails to explain why there is in this enormous number of - 61 -supposed interrogative sentences only one case of hu (for both.Kuan Hsietrch-u and Ch'en Meng-chia, all ming-tz'u are interrogative sentences). (Serruys, 1974:23). Since hu -f" can be used as a verb 'to summon, to call upon' in sentences such as: test / call upon / Lady Ching / millet / receive / harvest Chin 645 (If) we call upon Lady Ching to plant millet, we will receive a (good) harvest. , We might interpret the inscription ' l£l jjj^ l\ fejfi ^ ' as 'we shall make a sacrifice to Hsien Wu; HsUeh Wu (i.e. Teacher Wu) will call out (the order) (or: will be called out to)' (Serruys 1974:23).*"' The following tui-chen pair clearly shows that even when the word hu ^ appears in sentence final position, it may also function as the verb 'to summon, to call upon'. V a v in XL Ping 3(4) hui / Prince Pu / call upon / trap It should be Prince Pu.that we call upon to trap (animals). A <S_ n o ni Ping 3(5) should not / wei / Prince Pu / call upon .... It should not be Prince Pu.that.we call upon .... Putting too much emphasis on aligning O.B.I, syntax with classical Chinese while neglecting the internal evidence of the O.B.I, will produce incorrect interpretations. - 62 -The pattern 'non-nominal-negative + noun' may be used to illustrate the fact that in comparing the O.B.I, and classical Chinese, superficial similarities should not be used. In classical Chinese, nominal expressions can only be negated by the nominal negative fei or pu-wei j\\ . In cases where a 'noun' (defined by the semantic nature or quality of a word, not by its distribution)^ is preceded by a non-nominal negative such as p_u , the 'noun' will become verbalized. In the Lun Ytl ^flj? |^ , there are sentences such as: & * * (Lun YU HY 23/12/11) '...the prince be not prince, the minister not minister, the father not father, and the son not son.' (Legge, p.256) The second 'noun' of each clause is verbalized. In the case of O.B.I., we have sentences such as ) 'should not/one/ox'. The 'noun' after the negative wu ^) is not verbalized, i.e., the sentence cannot be interpreted either as "should not consider it an ox1 or 'should not make it into an ox', since verbalizing the 'noun' does not make sense in the context. In fact, the pattern 'non-nominal negative + noun' in the O.B.I, is different, in the underlying structure, from that of the classical Chinese. From the larger context, we can always supply a deleted verb between the non-nominal negative and the 'noun', for example, yu-offer / father / one / ox We should yu-offer to father one ox. Ping 52 (8) - 63 -__ m Pin£ 52 (9) should not / one / ox We should not (yu-offer to father) one ox. ,,j Ping 52 (10) ~ - r *f two / ox (We should yu-offer to father) two oxen. Ping 52 (11) should not / two / ox We should not (yu-offer to father) two oxen. Following the assumption of sentence parallelism (see p.69) in tui-chen ^ B pairs, the simplest solution is to take the sentence 1 ~~ /:f ' as having the verb deleted.*7 The deletion of a verb to give emphasis to other elements in the sentence is a practice well attested in Chinese. For example, # -f % ±. m m &. osse*** -37/5A/7) "I am one of Heaven's people who have first apprehended; I will take these principles and instruct this people in them. If I do not instruct them, who will do_ so?" (Legge, p.363) - 64 -/f t ± <^ * W * tl % & „ (Ibid.) "I have heard that I Yin sought an introduction to T'ang by the doctrines of Yao and Shun. I have not heard that he did so by his knowledge of cookery." (Legge, p.364) (Underlining the author's. The underlined words in translations show the omitted verbs in the originals. See also Chang Ping-ch'Uan (1965: Vol.2.1, p.460) and Ch'en Wei-chan j^. ^ (1980:196) for discussions on the practice of abbreviation in the O.B.I.) In modern Chinese, a similar phenomenon is observed. A response to the suggestion ' jang t*a ch' ih tien fan pa |^ ^ |fe /gfo />g_,let him eat some rice' may be 'pu yao fan, mien t'iao pa -Jj ^f^, /(\^ don't (give) him rice, (give) him some noodles'. - 65 -III. -THE SEMANTIC CONSIDERATIONS The O.B.I, language is very terse. We have not as yet discovered ,any sentence final particles or connectives. It remains undertermined whether this represents the true nature of the Shang language or is due to ellipsis in the notation of divinatory sentences, i.e., a kind of official short-hand. In contrast to its terseness, the O.B.I, language shows great variety in terms of sentence structure. To take a 'three-place' verb, liao jjj^ (burn-sacrifice) as an example, it can appear in the following positions: Ch'i test / burn-sacrifice / to / Ting / five / ox , We should burn-sacrifice five oxen to Ting. kuei-hai / crack / Ch'ueh / test / burn-sacrifice / (Shang)Chia three / ox / yu-offer / behead / ten / Ch'iang / ten / boar We 'should burn-sacrifice three oxen to (Shang) Chia, yu-offer one beheaded victim, ten Ch' iang-tribesmen and ten boars. Ho^- 162 to / (Shang)Chia / burn-sacrifice / three / small / penned-sheep dismember-sacrifice / three 75 (It is) to (Shang) Chia (that we should) burn-sacrifice three small penned-sheep and dismember-sacrifice three... - 66 -one / sheep / burn-sacrifice I L 7061 (It is) one sheep (that we should) burn-sacrifice. The verb liao sometimes precedes the indirect object(beneficiary) and sometimes is preceded by it. Its relationship with the indirect object Ch'en Wei-chan lists many examples showing the optionality of v_U (1980:177-rather than 'one sheep will burn-sacrifice' is obviously based on the semantic consideration that a sheep cannot be the agent of a sacrifice. By oxen to Shang Chia' but not 'burn-sacrifice Shang Chia and three oxen', since the ancestor Shang Chia, according to our knowledge, is not something that can be used as a sacrificial item. Semantic considerations of the above type lead us to interpret the following two inscriptions as semantically equivalent, i.e., verb + indirect object(beneficiary): (beneficiary) may or may not be indicated by the preposition yll 'f' . (Cf. the same token, 'burn-sacrifice three Ping 50 (3) We should exorcise to Father I. Ho 179 We should exorcise (to) Father I. Needless to say, to interpret these two inscriptions as equivalent on semantic grounds should be supported by other arguments: - 67 -1. There are obvious reasons to believe the preposition _y_U ^ is omitted in the second inscription since the practice of ellipsis is very common on the O.B.I. The following four inscriptions found on the same plastron illustrate this practice: ^ ff Ping 197 (5) come / chia-shen / yu-offer / to / T'ai Chia In the coming chia-shen day, we should yu-offer to T'ai Chai. next / ting-yu / yu-offer / to / Ancestor Ting On the next ting-yu day, we should yu-offer to Ancestor Ting. next / hsin-ch'ou / yu-offer / Ancestor Hsin-On the next hsin-ch'ou day, we should yu-offer (to) Ancestor Hsin. . next / i-ssu / yu-offer / Ancestor I . On the next i-ssu day, we should yu-offer (to) Ancestor I. On the entire plastron, there is no indication that the Ancestors in the former two inscriptions are beneficiaries while the ancestors in the latter two sentences are patients. - 68 -2. In the sentences where the pattern 'sacrificial verb + ancestral name' is incorporated, one never encounters a second personal figure which can be syntactically (e.g. sacrificial verb + ancestral name + y_U + personal name) or semantically (e.g. the personal name refers to a god or deity) interpreted as a beneficiary who might receive an ancestor as a sacrifice. Since sentences where the pattern 'sacrificial verb + animal noun(patient)' is incorporated often express an beneficiary (e.g. 'ij? god,*,<?) river god and ^jj). ancestor which may be introduced by the preposition vjl ^ ) , one is obliged to explain why, if an ancestral name is interpreted as a patient, there is never a beneficiary expressed in sentences where the pattern 'sacrificial verb + ancestral name' is incorporated. Unfortunately, what conditions the employment of y_U ^ is still problematic. Chou Fa-kao has observed that in the classics, indirect objects referring to people of high status are often introduced by the preposition ytl (1975:308). This principle is not applicable to the O.B.I, since all the dieties or ancestors presumably possess high status. And as inscriptions Ping 47 and Ho 179 show, even if the beneficiaries are identical, the occurrence of y_U -j"* is not predictable. - 69 -IV. THE ASSUMPTION OF SENTENCE PARALLELISM AND THE EXPLOITATION OF CH'ENG T'AO CSET) 1. The Assumption of Sentence Parallelism It is well known that the O.B.I, divinations frequently occur in pairs. To divine about whether 'it will rain or not' normally takes the following pattern 'it will rain' versus 'it will not rain' and these sen tences thus form a tui-chen pair. It is assumed that, in most cases, the grammatical structures of the two sentences of a tui-chen pair are identical. The notion of 'sentence parallelism' was initiated by Takashima (1973:288-305). In the article 'Subordinate Structure in Oracle Bone Inscriptions with Particular Reference to the Particle ch'i ^ ', Takashima has illustrated, with the following tui-chen pair, how the notion of 'sentence parallelism' can be applied in grammatical analysis: y* **. ts I* test / there is / come / start from / west there is not / perhaps(less desirable) / come / fit & west Takashima states, "If we were provided with only (.1), as is often the case in other collections, an unwary reader might analysis the sentence as having a syntactic break after lai %^ and interpreted it in the following way: "If there is an occasion (.for someone) to come, he will start from the west." This makes perfect sense and is in accordance with the general syntactic structure of the bone language and of classical Chinese. However, if we apply the same analysis to (2), we would expect the following interpretation, "if there is perhaps (i.e., we do not really wish it) no occasion (for someone) to come, he will start from the west." This, of Ping 94 (1) ti a start from Ping 94 (2) - 70 -course makes no sense at all and yet it does not violate the general syntactic structure of the bone language." (.1977:40-41) In other words, sentence (!) has two syntactically and semantically possible interpretations, i.e., Ca) 'If there is an occasion (for someone) to come, he will start from the west' and (b) 'There is an occasion (for someone) to come from the west'. But sentence (2) only allows one seman tically satisfying interpretat ion, i.e., 'There is no occasion (for someone) to come from the west'. It is on the assumption that the grammatical structures of a tui-chen pair are identical that we discard interpretation (a) of sentence (1). One may ask, is it possible to reject the assumption of 'sentence parallelism'? That is, to adopt interpretation (a) 'If there is an occasion (for someone) to come, he will start from the west' for sentence (1), and adopt the interpretation 'There is no occasion (for someone) to come from the west' for sentence (2). While possible, it is highly unlikely. Since the focus of sentence (a) is 'from which direction' the person will arrive, while the focus of sentence (2) is 'whether someone will arrive from the west' as a whole, if we adopt these two interpretations, we must provide evidence to explain why there is a 'shift of focal point' in these two sentences which belong to one single divinatory issue. It is because of 'sentence parallelism' that the following tui-chen pair is to be analysed as 'main clause + subordinate clause' although the first one, on purely syntactical and semantic grounds, could be analysed as 'subordinate (conditional) clause + main clause': ±t 1)« ftf?* ft* king / follow / Wang Ch'eng / make punitive expedition Hsia Wei / receive / abundant / assistance The king should follow Wang Ch'eng to make a punitive expedition against Hsia Wei, (for) he will receive abundant assistance. (But not 'If the king follows Wang Ch'eng to make a punitive expedition against Hsia Wei, he will receive abundant assistance.) king / should not / follow / Wang Ch'eng / make punitive -V f/t esjjt f ig_ ') ^_ expedition / Hsia Wei / not / perhaps / receive / assistance Ping 22 (2) The king should not follow Wang Ch'eng to make a punitive expedition against Hsia Wei, (for) he will not receive assitance. 2. The Exploitation of Ch'eng T'ab & -§\ (Set) The notion and the importance of Ch'eng T'ao was first introduced and elaborated systematically by Chang Ping-ch'Uan T|t-(I960). Ch'eng T'ao inscriptions refer to those positive and negative sentences which are divined on the same day, concerning the same issue, having similar meaning and having numerals (Jn.sU. shu ffi «^ ) in sequence. The notion of Ch'eng T'ao is of great significance in reconstructing elliptical sentences and thus may influence the analyses or interpretations of these sentences. For example, in the Ping Pien 34, there is a pair of tui-chen sentences: >Jhr 1^ ^ ii li li i-mao / crack / Ch'Ueh / test / king / preside / millet first _ 34 (5) test / king / should not / preside / millet first t 34 (6) At first sight, one would be to interpret them as 'The king should preside over the planting of millet' and 'The king should not preside over the planting of millet' respectively. However, on Ping Pien 35, we find the following pair: i-mao / crack / Ch'Ueh / test / king / preside / millet A, 3 approve second 35 (5) test / king / should not /preside / millet second 35 (6) (The interpretation of jL- (to preside over) is proposed by Serruys 1974:92). Judging from the positioning on the plastrons of these inscriptions and judging from other inscriptions which concur in this placement as well as the highly similar wording and identical date, there is no doubt that these two pairs belong to the same set. One may reasonably hypothesize that these two pairs, in the underlying level, are identical and it is only because the word ^ 'approve' is understood that it is omitted. If the pair on Ping Pien 35 can be interpreted as 'The king should preside over the planting of millet, (for) he will be approved' and 'The king should not preside oyer the planting of millet, (for he will not be approved)' respectively, then the pairs on Ping Pien 34 should also be interpreted in the same way, i.e., not as simple sentences as suggested above but as complex sentences. - 73 -CHAPTER TWO THE IDENTIFICATION AND CLASSIFICATION OF COMPOSITE SENTENCES WITH PARTICULAR REFERENCE TO THE 'CAUSE AND EFFECT' TYPE As mentioned in Chapter One, the study of O.B.I, grammar has drawn very little attention from scholars. In particular, the problem of composite sentence seems to have been ignored. Kuan's book lists six examples under the heading 'composite sentence' ~%%^ ^Qn *Q without any further sub-categor ization, and the 'discussion' of these sentences occupies only two lines. Ch'en Meng-chia deals with the problem of 'sentence-patterns' j^) in three pages without singling out the composite sentence for special attention of any kind. Nevertheless, a more accurate understanding of the O.B.I, language cannot be obtained if the logical relationship between the clauses of composite sentences is not touched upon. Though there are very few formal connectives, if any, in the O.B.I., there is no doubt that there are composite sentences. And as a matter of fact the paucity of connectives in composite sentence is one of the characteristics of the Chinese language. In the Mencius, we find sentences such as the following: 1/1A/3) If the seasons of husbandry be not interferred with, the grain will be more than can be eaten. (Legge, p.130) Although there is no 'visible' marker, this string can only be understood as a 'composite' sentence once the context is taken into account. In. a strict sense, no matter how closely two clauses are related in meaning, they cannot - 74 -be termed as clauses of a 'composite' sentence if there is no marker. A marker may either be segmental, e.g. ^ 'if'» or suprasegmental, e.g. pause and/or pitch. At the present time, it is no longer possible to discover whether there were any suprasegmental markers imposed on sentences such as the one cited above. We can only assume that there were suprasegmental markers not recorded in the written language and interpret such sentences as parataxis. In instead of formal markers, it is the contextual meaning which determines logical relationship between two particular clauses. The classification of composite sentences varies from one scholar to another. For example, Wang Li lists twelve sub-classifications under the heading 'composite sentence' (1946:116-117); Chou Fa-kao gives six (1961:198). (condition) (concession) (cause and effect) (simultaneous and successive) (adversative) (cumulative) Although the number of sub-classifications differs from scholar to scholar, discrepancies in the actual analysis of composite sentences do not often occur. In this study, Chou Fa-kao's classification has been adopted for its simplicity. Cases requiring a subtler analysis will be disccused individually as they occur. The term 'subordinate composite sentences' refers to the first five kinds of sentence described above. However, due probably to the particular character of the Shang divinations, 'concessive' sentences (i.e., sentences showing a logical relationship marked by connectives such as sui ^f^'even' in classical Chinese) and 'adversative' sentences (i.e., sentences showing (1) i & t? (2) (4) #) io (6) * tf & - 75 -the logical relationship which is marked by connectives such as (jari)erh (^f>.^ \5y 'however' in classical Chinese) do not seem to exist as major categories in the O.B.I. There are some sentences, which, when considered separately, are open either to the 'concessive' or 'adversative' analysis. For example, king / yu-offer / captive pair / not / favorable Even if the king yu-offers a pair of captives, it will not be favorable. , £ £ «i hi _ P ft 4A king / perhaps / go / chase / -game(?) I at I Kuei^-not / perhaps / capture Ping 216 (6) The king will perhaps go to chase ^ -game(?) at Kuei, however, he will perhaps not capture (them). Or: The king will perhaps go to chase ^? -game(?) and kuei-game (?), however, he will perhaps not capture them. (For the interpretation of the words ' ^ ', see Chapter Three, footnote no. 14.) . Syntactically and contextually, the above analyses are, possible. Nevertheless, such analyses cannot be applied to the respective positive counterparts of these two sentences: kuei-wei / crack / HsUan / test / king yu-offer / captive pair / favorable - 76 -Even if the king yu-offers a pair of captives, it will be favorable, king / perhaps / go / chase / ^ -game(?) / at kuei / capture The king will perhaps go to chase ^ -game(?) at Kuei, however, he will perhaps capture (them). Or: The king will perhaps go to chase -game(?) and kuei-game(?), however, he will perhaps capture (them). Obviously, the last two interpretations do not make much sense. If we adopt the notion of sentence parallelism in our analyses, the 'concessive' or 'adversative' interpretations of the negative counterparts must also be rejected. We may, on the other hand, interpret all four of these inscriptions as conditional sentences. That is, Ping 53 (2) 'If the king yu-offers a pair of captives, it will not be favorable. 53 (1) 'If the king yu-offers a pair of captives, it will be favorable. Ping 216 (.6) 'If the king perhaps goes to chase ffi -game(?) at Kuei, he will perhaps not capture (them). Or:'If the king perhaps goes to chase ^ -game(?) and kuei-game(?), he will perhaps not capture (them). 216 (5) 'If the king perhaps goes to chase ^ -game(?) at Kuei, he will perhaps capture (them). Or:'If the king perhaps goes to chase ffi -game(?) and kuei-game(?), he will perhaps capture (them). Inscriptions Ping 139 (.1) (2), 1-57 .'CD C2), 159 (.8) (.9) and 189 (9) (.10) are also subject to the above analyses. I. THE CRITERIA FOR CLASSIFYING COMPOSITE SENTENCES Since there are no formal markers (connectives) in the O.B.I. and the presumed pause, pitch or intonation is not shown in the inscription, the classification of individual composite sentences is highly problematic. Concerning the omission of connectives in Chinese, Wang Li states, "The fixed position (of the two clauses in) the •'subordinate + main' sentence is the major reason why connectives may not be employed. ... In Chinese when two sentence-forms are joined together, although there are no connectives, we are still able to tell that the former one incorporates a meaning such as 'although', 'even if, 'if, 'if, 'if» 'since' or 'because of, this is because the subordinate clause must be placed initially." i.\t q < ft w @ * «k * % * f mlS> .... $%&t®/®*!* ^'#?it^it i - € % K It is rather difficult to understand Wang's argument. The fixed (.1946:125) - 78 position of clauses can only assist us in determining which clause is sub ordinate and which is main. The precise logical relationship between the clauses (e.g. 'conditional') is not shown by this positioning. It appears that the only criterion for determining this logical relationship is the contextual meaning. In classical Chinese, not to mention more modern forms of the language, a larger context relevant to the sentence in question is frequently available, and thus there are not many cases of possible con troversy in interpreting the logical relationship between two clauses. In the O.B.I., however, the contextual relationship between sentences is not always clear and 'larger context' is only, a vague term referring to sentences appearing on the same plastron or on the related plastrons In some cases, the contextual relationship between sentences is so obvious that we may consider them a divinatory group. ). Ping 1 (9) (Ch 'tteh) will capture the Fou-(tribesmen). Ch'ueh will not capture the Fou-(tribesmen). Our envoy will harm Fou. Ping 1 (14) Our envoy will not perhaps harm Fou. - 79 -Although the dates of divination and the diviners mentioned in the above two pairs of sentences are different, we can still consider each pair to be related contextually on the basis of the identical object ID Fou preceded by the verb ^ 'capture' and 'harm' both seem to be related to military activities on this particular plastron. In contrast to the above pairs, there are some inscriptions whose inter-relationship is very obscure, though they appear on the same plastron: :hi-wei / crack / Cheng / test / Wang Hai chd harm / we \ Wang Hai will harm us. test / Wang Hai / not / we / harm ^Wang Hai will not harm us. test / we / perhaps / have / disaster We will perhaps have disaster. test / we / have no / disaster ^We will have no disaster. chi-wei / crack / Ch ueh / test / we / at < Chih / enter / station We should enter at Chih and station (there) test / should not / at / Chih / station Ping 3 (1) Ping 3 (2) Ping 3 (11) Ping 3 (12) Ping 3 (19) Ping 3 (20) We should not station at Chih. - 80 -Whether the divinations about ' % 1$, ' are related to the action ' -j-" £^f, Jj, ' is something we cannot be sure of. Likewise, it is uncertain whether ' "£ ^ ^ 'is the cause of ' ^ rf%$aj Undeniably,) subjective interpretation is sometimes the only means left to us in positing the contextual relationship between sentences. Nevertheless, in view of the fact that there is no other way, we must exploit this 'larger context' as fully as possible on the assumption that the sentences appearing on the same plastron are related in one way or another. Without this assumption, a lot of information, including that concerning the determination of the type of composite sentence, cannot be recovered. In addition to the meaning and context, the 'general pattern' jj£ of O.B.I, and the practice of ellipsis have to be taken into consideration also. Instead of going into theoretical discussions, the author presents the following example to illustrate an actual application of the points men tioned above in the analysis of composite sentences. hsin-yu / crack / Ch'Ueh / test / this / season ±* ))u \* f ji*; ft* king / follow / Wang Ch'eng / attack / Hsia Wei receive / abundant / assistance Ping 20 (1) Interpretation A: This season the king should follow Wang Ch'eng to attack Hsia Wei, (for) he will receive abundant assistance. Interpretation B: If, this season the king follows Wang Ch'eng to attack Hsia Wei, he will receive abundant assistance. - 81 -hsin-yu / crack / Ch'Ueh / test / this / season king / should not / follow / Wang Ch'eng / attack Hsia Wei / not / perhaps / receive / abundant / assistance Ping 20 (2) Interpretation A: This season the king should not follow Wang Ch'eng to attack Hsia Wei, (for) he will not perhaps receive abundant assistance. Interpretation B: If the king does not , he will not Concerning the relationship between two clauses such as those of Ping 20 (2), different analyses have been proposed. :ives as 1. Chou Hung-hsiang J^} >ftfj fyjinterprets the two negati forming a 'double negative' construction, the negatives cancelling each other out; thus, sentence (2) has the same cognitive meaning as sentence (1), i.e., both sentences are positive rather than positive versus negative. (1969:77) Unfortunately, Chou has not explicated what he means by 'double negative' ^ tf\ ^- Q %_ . It seems that he treats the two negatives wu and fu $p in Ping 20 (.2) in the same way as he treats the 'double negatives' wang ~{Z- and p_u ^- in sentences such as ' J_ £. rf\ % 1 (the king has no uneasiness). However, the negatives wu 'fl] and fu $p in sentence Ping 20 (2) occur in different clauses while wang and pu rf- do not. The relationship between wang JZ- and pu rf> cannot be the same as that between wu fyj] and fu . Thus, Chou's interpretation is difficult to accept. - 82 -2. Serruys analyses such sentences as 'main clause + subordinate clause' and supplies a paraphrase which has a 'subordinate (conditional) + main' structure. "Test (the proposition): we ought not attack the Hu (country) .... (for) God might not give us (lit. make receive) his assistance." (i.e., "If we do not God will not ....) (1974:44) Do not burn (any victim at the stake), then we will not have ensuing rain. Burn (victims at the stake), then we will have ensuing rain." The prohibitive "don't" and its counterpart "do burn" are often used in the function of conditional: "If we do not burn " and "if we burn " This is comparable to similar structures in other language, e.g. classical Greek usage of negative me (cf. Hansjakob Seiler, "Abstract Structures for moods in Greek" Language, Journal  of the Linguistic Society of America, Vol. 47, 1. 1971, p.81): "If one looks at -a.number of conditional sentences it seems odd that the protasis should invariably be negated by means of the prohibitive negation." (1974:27) As shown in the translations of the sentence ' ^ ^) ^£ <5 ', Serruys understands the structure 'if ... then ' as equivalent to 'should , for '. Precisely as Takashima has pointed out, these two structures cannot be equated: It seems to me that "we ought not proceed to hunt, for it will not be caught" is not at all equivalent to "if we do not proceed, it will not be caught." That is, the illocutionary force of the former seems to suggest "so, don't proceed to hunt," whereas the illocutionary force of the latter suggest, "so, do proceed to hunt." (1977:53) - 83 -In any case, the translations provided in parentheses by Serruys reflect a structural analysis identical to Takashima's interpretation A which the author considers a-correct reading. This problem is dealt with in the following sections. type: 3. Takashima has proposed two interpretations for sentences of this A. main + subordinate(causal) e.g. $• * i to AK & % * * 'The king must not follow Wang Ch'eng to attack Hsia Wei this spring, (for) (he) will not perhaps receive abundant assistance'. (1973:292) B. subordinate(conditional) + main e.8. 'If the king does not follow Wang Ch'eng to attack Hsia Wei this spring, he will not perhaps (less desirable) receive abundant assistance. (1977:44) Interpretation A is a simple and natural analysis based on the context. The word wu ^) , frequently functions as a prohibitive negative both in the O.B.I, and in the classics, so it is reasonable to interpret the clause » f ^ « as 'the king should not follow Wang Ch'eng to attack Hsia Wei'; and since the clause ' (will not perhaps receive abundant assistance) represents an undesirable situation, it is also reasonable to interpret it as the cause for prohibiting the king from following Wang Ch'eng to attack Hsia Wei. On the other hand, the above analysis which places a subordinate clause after a main clause is somewhat unusual in terms of the predominant - 84 -structure of classical Chinese. Takashima points out that 'from the stand point of the general syntactic structure of Chinese, the apodosis usually follows the protasis' (1977:54). Thus, interpretation B is proposed as another possibility. In addition to the example in classical Greek, evidence is cited from English and Japanese to demonstrate that an imperative/prohibitive sentence-form may function as the protasis of a conditional (1977:51-52). From these examples, there is no doubt that in many languages, imperative/prohibitive sentence-forms may at times have this function. This interpretation may be applied to certain classical Chinese sentences as well. For example, In this juncture, it appears that each interpretation, A and B, has its merits and can be supported on different grounds. However, the fact that both of these two interpretations are possible does not mean that these two logical relationships are simultaneously incorporated in one single sen tence (the illocutionary force of these two interpretations are opposite to each other, as Takashima has pointed out). There can be only one logical relationship between the two clauses of the sentences in question. The task before us is to test which of these two interpretations is more appropriate. The present author holds the opinion that interpretation A is more likely to be the correct one for the following reasons: (Mencius HY 1/1A/3) 'Let there not be taken away the time that is proper for the cultivation of the farm with its hundred mau, and the family of several mouths that is supported by it shall not suffer from hunger.' (Legge, p.131) - 85 -(1) The Symmetry of Tul-Chen Pairs It is a well-known fact that members of most of the tui-chen pairs have a sense opposite to each other, i.e., to have ( ^ ) vs not to have ( ) it is ( J^i ) vs it is not ( ^ ]i ) will (uncontrollable vs will not verb) uncontrollable verb) to do (controllable vs not to do ( ^ + control-verb) lable verb) In other words, each member of a tui-chen pair states a possibility of which the other member is the negative counterpart and each pair covers all the logical possibilities. If we adopt interpretation B, that is: "if the king follows Wang Ch'eng to attack Hsia Wei this spring, he will receive abundant assistance." "if the king does not follow Wang Ch'eng to attack Hsia Wei this spring,, he will not perhaps (=less desirable) receive abundant assistance." The illocutionary force of these two sentences turns out to be the same the king should follow Wang Ch'eng to attack Hsia Wei this season. It goes without saying that this interpretation violates the general dualism which characterizes O.B.I, divinations; in other words, this pair of tui-chen covers only one possibility by means of two equivalent expressions. One may argue that this tui-chen pair should not be treated indepen dently. It has to be investigated together with other tui-chen pairs appearing on the same plastron: - 86 -hsin-yu / crack / Ch'Ueh / test / king / hui Chia / follow Ping 20 (5) J It should be Chia (rather than anybody else) that the king follows. hsin-yu / crack / Ch'Ueh / test / king / should not wei / Chia / follow Ping 20 (6) It should not be Chia that the king follows. In the final analysis, the main concern of this divinatory group (Ping 20 (1) (2) (5) (6)) is 'whom the king should follow, Wang Ch'eng or (Chih) Chia', so the divination was made to determine the choice between these two generals. (Theoretically, there are four possibilities: i. to follow Wang Ch'eng; ii. to follow Chih Chia; iii. not to follow either of them; iv. to follow both of them at different times.) Even if we accept the above explanation, to interpret Ping 20 (1) (2) as having the same illocutionary force requires the assumption that the Shang king has already made up his mind or, at least, has a strong inclination to follow Wang Ch'eng rather than Chih Chia. We must admit that such a possibility exists if we adopt Keightley's hypothesis (1972) that a ming-tz'u, in some cases, should be understood as a prayer or incantation. However, as illustrated in the following passages, Ping 21 (1) and (2) would better be understood as divinations soliciting instructions from a supernatural power to determine whom the king should follow, rather than prayers or incantations made to seek confirmation of a priori decision on the proper course of action - 87 -to follow. If, on the other hand, we look at the three pairs of tui-chen as a unit, interpretation B destroys the symmetry of the whole. See Figure I. The illocutionary force of sentences (3) and (4), (5) and (6) are contrary to each other. It is unlikely that (1) and (2) have similar illocutionary force. It should be noted that these three pairs of tui-chen sentences are symmetrically positioned on the plastron. One more example, Ping 22, may further demonstrate the strange results obtained when one analyses sentences of the 1 ^7 , &I ^ 1 type as conditional. See Figure II. If we were to say that sentences Ping 22 (1) and (2) have similar illocutionary force, based on the assumption that the king has formed a prejudgement or is inclined to follow Wang Ch'eng, it would be difficult to account for sentences (3) and (4) which have illocutionary force contrary to each other. In terms of placement, meaning and illocutionary force, sentences (3) and (4), (11) and (12), (13) and (14) are opposite. It would be highly unlikely for sentences (1) and (2), which are obviously related to the other three pairs, to have identical illocutionary force while retaining opposing placement on the plastron. . In general terms, divination is usually understood as a means of foretelling future events or of requiring super-natural guidance for one's actions; in either case, it is generally presupposed that the person initiating the divination has no knowledge of what will happen and is ingenuous in his seeking of the oracle's guidance. To interpret the 'ID ^ / ......' type sentence as a prohibitive functioning as conditional which carries a strong illocutionary force and which at the same time turns out to - 87 a -If the king does not follow Wang Ch'eng to attack Hsia Wei this ^ \ 3i-season, he will no perhaps receive . v=> -a. abundant assistant lC-> iff *i AA t3 2-4 A A The king should not follow Chih Chia. AA Jjfc 6 It should not be Chih Chia that the king follows. If the king follows JVang Ch'eng to attack Hsia Wei Nth is season, he Ti\will receive q: \ abundant j3f \assistance. AA. It should be Chih Chia that the king follows. FIGURE I (Ping 20) -87 b -If the king does not follow Wang Ch'eng to attack Hsia Wei, he will / jjp AT" not receive assistance AA-A/. The king should not follow xn jj] Wang Ch'eng. ^£ -7 ufj AA 13 If the king follows Wang Ch'eng to attack sia Wei, he will receive abundant assistance. AA 4, R' 1 ±-The king should follow Wang Ch'eng. It should not be Chia that the king follows. It should be Chia (that the king) follows. The king should not follow Chih Chia to attack Pa. AA ^7/ /3 AA -r ft, It should be Chih Chia that the king follows to attack (Pa). FIGURE II (Ping 22) - 88 -have an illocutionary force similar to its positive counterpart ' % ^K %- $t. ^ Tv JZJ » ^k. 'i? ' would overload this tui-chen pair with illocutionary force. (It appears difficult to discover an illocutionary force in the pair ' ;£ AA }Jz J^' 1 and ' 'f/J M >>t ' strong enough to counterbalance this reading of our example.) More importantly, interpretation B seems contrary to our basic assumption about divination. One may argue that the Shang king had already decided to follow Wang Ch'eng before the divination was performed, so that what was needed was merely confirmation from the supernatural world. If this were the case, why did the diviner proceed to divine about whether the king should follow general Chih Chia or not? The concurrence of divinations about following Wang Ch'eng with that concerning Chih Chia shows that the purpose of these divinations is to choose between these two generals rather than to seek confirmation for following Wang Ch'eng. (2) The Elliptical Sentences As mentioned in Chapter One, the practice of ellipsis is very sty . common in the O.B.I. While the omission of the ch ien^tzu ff<J , e.g., ^f7 h f^K. |( ^s not at a±1 rare, there are also cases where an entire clause is abbreviated, for example: (•Tft» U %f W* 7f e&ffl ting-yu / crack / Ch'eng / test / call upon / Pu ni-rice / at / Tzu / receive / abundant / grain If we call upon Pu to (plant) the ni-rice at Tzu, we will | receive abundant grain.1.TJ 3212 - 89 -\>31$* u H$ HI yA ting-yu / crack / Cheng / test / not / perhaps ting-yu receive / abundant / grain I "Z^ 3212 We will not perhaps receive abundant grain.' < The conditional clause of the negative counterpart has been abbreviated. There are even some inscriptions which illustrate the process of ellipsis in stages: rffik k^ &k:>?':kl&>%' H*. jivi test / king / hui / Chih Chia / follow / attack / Pa statelet / ti-god / give / we / assistance I L> 3787 It should be Chih Chia that the king follows to attack the Pa statelet, (for) ti-god will give us assistance. king / should not / wei / Chih Chia / follow / attack Ji fi "M* &a y\ «Mt) Pa / statelet / ti-god / not / we / perhaps / give / assistance It should not be Chih Chia that the king follows to attack the Pa statelet, (for) ti-god will not give us assistance. I & 3787 test / king / follow / Chia / attack / Pa / ti-god fit *l give / assistance Ping 25 (1) The king should follow (Chih) Chia to attack the Pa statelet, (for) ti-god will give us assistance. test / king / should not / follow / Chia / attack / Pa Ping 25 (2) ^ The king should not follow (Chih) Chia to attack the Pa statelet, - 90 -(for ti-god will not give us assistance), test / king / follow / Chih Chia / attack / Pa Ping 26 /CD The king should follow Chih Chia to attack the Pa statelet, (for ti-god will give us assistance). tl £ HO W A< ^ &4 \ It rt & king / should not / follow / Chih Chia / attack / Pa Ping 26 (2) The king should not follow Chih Chia to attack the Pa statelet, V (for ti-god will not give us assistance). From the full version: * to £ & £"4$ * fb*'5L^ we proceed to a partially abbreviated pair: £ ^ ^ «fr i£ * and eventually reach the most abbreviated form: \ $ kk -;Jt % £ to £> Here the process of progressive abbreviation is very clear. With this practice in mind, we may, on the basis of the larger context, interpret Ping 22 (3) (4) (See Figure II) ' :£ AA ^ ' and as the understood, abbreviated versions of Ping 22 (1) (2) ' £ AA 4 * , '1 * *• and • £ ^ ^> $ ^ f ^ ^ >L' resPectively. If this is justified, the reason militating against interpreting the sentence ' ^ ^ ^ ^ ^ 7?j • ^ — «t^ Xj, as a 'subordinate + main' structure becomes clear. In the abbreviated form, the remaining clause is ' ^) J^l. ». if we treat this as a subordinate - 91 -(conditional) clause and translate it as 'if the king does not follow Wang Ch'eng' with positive counterpart 'if the king follows Wang Ch'eng', it does not constitute a meaningful charging statement. However, when analysed as main clauses, with the understood subordinate (causal) clause abbreviated, i.e., 'the king should not follow Wang Ch'eng' and 'the king should follow Wang Ch'eng' (clauses 'for he will not receive abundant assistance' and 'for he will receive abundant assistance' abbreviated), they do constitute meaningful charging statements. In support of interpretation B, the author once postulated the following argument: Although the clauses ' % A A *jjf and ' £ 77/ XA *§_ jfcl function as 'conditionals', if we take them at face value, they display the structure of prohibitions. Thus the sentences ' J AA $t ^\ ft-J » ^ <£ ' and ' £ 'fJT AA % |t f /% , % % ^' should be translated as 'King! Do follow Wang Ch'eng to attack Hsia Wei (and) you will receive abundant assistance' and 'King! Don't follow Wang Ch'eng to attack 3 Hsia Wei (and) you will not receive abundant assistance' respectively. In this case, the abbreviated sentences ' A A ^ ' ANC* ' £• ^ ^A ^ 1 can be translated as 'King! Do follow Wang Ch'eng' and King! Don't follow Wang Ch'eng' respectively and thus do constitute meaningful charging statements. However, in adopting such an interpretation, the illocutionary force of the sentence ' £ jjj AA 4Ji /t> » ^ J% ^ ^ ' ±s °PP°site to that of ' 3E. ^ AA ' t*ie f°rmer urges the king to follow Wang Ch'eng (an analysis of illocutionary force based on interpretation B) while the latter urges the king not to follow Wang Ch'eng. Unless we are - 92 -willing to deny the obvious fact that the sentence 1 3- ^ Iff is'.the abbreviated form of ' % 'ij] AA % $L 4$ T /£, , 4 %L % X-1, such a change of illocutionary force is unlikely. (3) The Analysis 'Main + Subordinate' Motivated by Semantic Considerations in the O.B.I. While it is a general pattern of classical Chinese that apodoses follow protases, there is evidence showing that the opposite case is also possible in the O.B.I. The following inscriptions exemplify the structure of main clauses preceding subordinate clauses: i* W^'rn) fflffl Ux hsin / king / should not / hunt / perhaps / rain On the hsin day, the king should not hunt, (for) it will perhaps rain. Ts'ui ffi 1008 ' -" .- ' jen / king / should not / hunt / perhaps / rain On the jen day, the king should not hunt, (for) it will 'vperhaps rain. Ts'ui 1008 As mentioned above, there are no segmental markers (i.e. connectives) between clauses in the O.B.I., thus the logical relationship between the clauses can only be deduced from the context. The author has tried to analyse these sentences in other ways, for example, i. 'On the hsin/jen day, if the king does not hunt, it will perhaps rain.' - 93 -ii. 'On the hsin/jen day, the king should not hunt and it will perhaps rain.' iii. 'On the hsin/jen day, although the king should not hunt, it will perhaps rain.' Interpretations (i) and (ii) require the assumption that whether one hunts or not has an effect on the falling of rain, obviously a very remote possi bility. Interpretation (iii) does not make any sense in terms of our under standing of the relationship between hunting and rain. It appears that the most appropriate, if not the only possible, interpretation is '... the king should not hunt, (for) it will perhaps rain'. Two other tui-chen pairs also favor the 'cause and effect' analysis. should not / burn-sacrifice / have no / rain We should not perform a burn-sacrifice, (for) there will be no rain, ^* $ 1 ft r tr % *fcTTaf 189 perhaps / burn-sacrifice / at / Hslleh / have/big / rain We should perhaps perform a burn-sacrifice at Hstleh (or, to the god of snow), (for) there will be heavy rain. h ft &* fa ?m hui / Sang / field / inspect / not / rain Menzies 1900 It is the Sang fields which we or the king should inspect, (for) it will not rain (over there). should not / inspect / Sang / field / perhaps / rain We or the king should not inspect the Sang fields, (for) it will perhaps rain (over there). (Examples of this type, though not in the tui-chen format, can be found on the Sorui, p.293.3.) - 94 -There is a difference of degree in terms of the compelling force expressed in the clausal causes. For instance, in the sentence ', the relationship between the clauses expresses a condition, rather than an explicit reason for the king to inspect the Sang fields. On the other hand, the negative counterpart 1 jfy ^/L.XiO , 1 seems to express a compelling reason (see Chapter Three, p.183- 186), if not one so strong as in the sentence '/Cs| £ U ^ ? & > $ J; £ ^ £ ' ' In all cases, however, there is a logical relationship 'cause and effect' (in the linguistic sense) in all these sentences. All six of these examples are late period inscriptions (the inscrip-tion '^fau^ jfefylfe, %^%^^K_X is a first period one) . It is unlikely that, in the development of the Chinese language, the structure 'main + subordinate' unknown in the first period, would appear in the later periods and then disappear again by the classical period. Instead, it is more reasonable to hypothesize that the structure 'main + subordinate' existed in the O.B.I, period, but gradually disappeared or took another form 4 in the course of the historical development of Chinese. To recapitulate, it is for the following three considerations that the author chooses not to interpret sentences having the structure ' ^ ..., ' as conditional sentences: (1) Such an interpretation violates the general pattern of positive versus negative in the ming-tz'u. (2) From the standpoint of the practice of ellipsis, clauses with wu ^} are. usually retained in the abbreviated forms. It would sound very odd if these were interpreted as protases, - 95 -e.g., ^ Ml *^^'*if the king does not follow Wang Cheng'. (3) The context of certain inscriptions shows that the structure of 'main + subordinate' exists in the O.B.I. Taking into consideration all the points discussed above, the author proposes that interpretation A is preferable to interpretation B. The former fits the general pattern of O.B.I, tui-chen pairs and avoids the pro blem arising from the latter. In addition, the analysis'main+subordinate' is supported by sentences such as ' ^ ^ (j^k) )^3? , life) ' of which semantic consideration makes the interpretation 'we should not ..., (for) . ..' the most appropriate. Keightley also analyses this type of sentences as 'main +'subordinate' in his translations (1978:78; see also p.66, fn.44). This season, the king should follow Wang Ch'eng to attack the Hsia Wei, (for if he does, we) will receive assistance in this case. This season, the king should not follow Wang Ch'eng to attack the Hsia Wei, (for if he does, we) will not perhaps receive assistance in this case. The cognitive meaning of Keightley's translations is very similar to that of interpretation A. And yet, as Takashima has pointed out in private communication to David Nivison, there is no basis to supply a clause 'for if he does' to the original sentence. Keightley's translations, though they read more smoothly, may blur the distinction between different types of sentence. - 96 -II. THE POSITION OF THE NEGATIVE WU pj IN RELATION TO THE ANALYSIS OF SENTENCES It has been illustrated in Chapter One that the assumption of 'sentence parallelism' has great significance in the correct understanding of sentence structure (Chapter I, p.69). Positive/negative alternation is characteristic of a parallel tui-chen pair and the position of the negative particle frequently functions as a 'mark' of a major syntactic unit. Again we use the example cited by Takashima (1977:44) to illustrate this phenomenon: 4 $k Q A ^ 94 (1) 'there is an occasion (for someone) to come from the west' -t Jt ^ tJ *b Ping. 94 (2) 'there is no occasion (for someone) to come from the west' The reason we cannot interpret as 'if there is an occasion (for someone) to come, he will start from the west' is because the initially placed wang in the negative counterpart makes nonsense of a parallel interpretation of {$&, i.e., 'if there is not an occasion (for someone) to come, he will start from the west'. While this interpretation of the syntactic function of negative particles is valid in general, there are some exceptions. These exceptions are due to the peculiar linguistic behaviour of wu 4^ , the negative which occurs most frequently in the O.B.I. In the following section, the peculiarities of wu ^ will be discussed in relation to the analysis of O.B.I, sentences. - 97 -As mentioned in Chapter One, p.62-64, in the classics, the negative wu ^£7 has to be placed in front of a verb and this is the characteristic feature, among others, which distinguishes it from the nominal negative fei In the O.B.I., however, the position of wu ^ is more flexible; it can be placed in front of a noun. For more examples: test / burn-sacrifice I to I Wang Hai / ten / ox We should' burn-sacrifice ten oxen to Wang Hai. ^ Ping 112 (8) k ft* uy* test / should not / ten / ox We should not (burn-sacrifice) ten oxen. Ping 112 (9) f test / call upon / Ch'Ueh / yu-sacrifice / to if) i 3t River(god) / fifty / ox < 1 We should call upon Ch'Ueh to yu-sacrifice to the River(god) fifty oxen. Ping 117 (5) should not / fifty / ox / yu-sacrifice / to / River (god) V We should not (call upon Ch'Ueh) to yu-sacrifice the River(god) fifty oxen. Ping 117 (6) Although, in the above examples, the negative wu immediately precedes nouns phrases, it has not been interpreted as a nominal negative. We rather describe these negative sentences as having verbs which have been elided or shifted as shown in the translations. Such an elision or shifting is a mechanism triggered by the intention of emphasizing the direct objects (sacrificial victims). - 98 -In his article 'On Negation with Fei in Classical Chinese', Sian L. Yen states, '"John didn't buy the book." If the stress falls on John, it means that, though John did not buy the book, someone else did. In other words, what is specifically negated is John On the other hand, if the stress falls on the word book, it implies that John actually bought something although what he bought was not the book in question.' (1971:409) whether there was stress in the O.B.I, language cannot now be determined. What is clear from the material accessible to us is that the position of the negative wu serves the function of distinguishing the focus of negation. Similar to the examples given by Yen, a sentence '* ' is open to two interpretations: a. It is not a burn-sacrifice that we should perform in presenting the ten oxen (i.e. it is some other rite). b. It is not ten oxen that we should present in performing the burn-sacrifice (i.e. it is some other animal or a different number of animals). In the O.B.I, language, the deletion or the shifting of the verb, which allows the negative wu ^ to appear immediately before the noun, has the function of disambiguating the focus of a divination. The sentence ' ^H) ~f" ^ ' is thus open to one interpretation only: 'It is not ten oxen that '. The larger context in which the sentence ' -J* ^ ' appears supports our theory. Together with the sentences ' j^f,-J* J ^ *f ty ' and ' ty) -j/ '» the following tui-chen pair appears on the same plastron: - 99 -test / burn-sacrifice / ten / ox We should burn-sacrifice ten oxen. Ping 112 (11) test / should not / specifically / three / ox Ping 112 (10) ^ We should not specifically (burn-sacrifice) three oxen. Obviously the focus of this group of divinations lies on the number of sacrificial animals.5 Similarly, on Ping 117 where the sentences 1 fy^. "j" ' appear, we find: thirty / ox / bring7 yu-sacritice / River (god) / thirty Wo / woman We should yu-sacrifice to River(god) thirty oxen and bring (as sacrifice) a woman of the Wo tribe. Ping 117 (8) With these examples in mind, we can proceed to analyse the following sentence: test / Yung 7 cut grassV at / Ch'iu Yung should cut grass at Ch'iu. Ping 141 (13) test / Yung / cut grass / should not / at / Ch'iu Ping 141 (14) ^Yung should not cut grass at Ch'iu. Yung / cut grass / at / Ku Yung should cut grass at Ku. Ping 396 (6) >. should not / at / Ku1 Ping 396 (7) (Yung)should not (cut grass) at Ku. - 100 -test / Kung / cut grass?/ should not / at/ % (?) Ping 413 (21) I a? Kung should not cut grass at $(?). r i. %$ NM? i-ssu / crack / Chung / test / Kung cut grass / at/ (?) Ping 413 (22) ^ Kung should cut grass at •^§r (?). As reflected in the translations, the above sentences are best analysed as simple sentences. Such an analysis, in fact, goes contrary to the general observation that the negative marks a major syntactic break. One naturally wonders why these sentences cannot be analysed as composite, e.g., 'if Yung cuts grass, he should / should not go to Ch'iu'. This latter analysis, however, requires that ytt ^ be treated as a full verb meaning'go'. This interpretation was, in fact, formerly adopted by the author, but has been discarded. The main reason for discarding such an analysis.is that yU ~f is not a full verb. Since the meaning and the function of yU are crucial to the understanding and analysis of this type of sentence, they deserve meticulous study. The following section is an excursion on the presumed full verb usage of yjU -jp in five kinds of materials where yjl frequently appears: namely, the O.B.I., the bronze inscriptions, the Shang Shu, the Shih Ching and the Tso Chuan. The glossing of yU ^ as ' ^{fa to go' is found in the Mao Chuan ^ 'fy (SSCCS, Shih Ching, Chuan 1.2, p. 15) and Cheng Chienlfj* \- (ibid., ChUan 4.2, p.9). That y_U J' has a verbal sense is recognized by many - 101 -scholars; among them, Karlgren in his translation of Shih Ching 'The Book of Odes' (1950:3), Takashima in the Negatives in the King Wu-ting Bone  Inscriptions (.1973:156), Yang Shu-ta ln his Chi Wei Chti Chia Wen Shuo (1974:12) and Han Yao-lung || ?f in his article 'Chia Ku Pu Tz'u Chung YU Tzu Yung Fa T'an Chiu Thus, ^ % f $ is glossed as ' % % & dji 'in the Mao Shih  Cheng I 1^ (SSCCS, Chuan 1.2, p.2). "The yellow birds go flying." (Karlgren 1950:3) However, as Chou Fa-kao points out, there are cases where the interpretation to go' is not applicable (1962:251). For example, 'fjL 9 ^ QlL (Shih Ching HY 28/137/3) "The morning being good for the excursion." (Legge, p.207) "An auspicious morning they proceed." (Karlgren 1950:88) £ ^ *+ (Shih Ching HY 38/177/1,2) "The king had ordered the expedition." (Legge, p.281) "The king sent out a war expedition." (Karlgren 1950:120) (The Mao Chuan and Cheng Chien do not gloss ' ^ ' as ' ' in these two 1 • • x 7 lines.) Ch'U Wan-li $ %) % renders the word yjl 5" in the line tj J ^ as tsai fa , yU fei -J- ^ means tsai fei rfj- jfjg (1967:3). Chtt's rendition does not fit this particular line very well, since the following line reads ' -f ' . ^ Nevertheless, to interpret yjl ~f as the modern co-verb tsai (to be distinguished from the full verb tsai ^ in both classical and modern Chinese) which has the function of showing an action in process - 102 -can be defended in other cases, (e.g. ' ,J- f /fj 1 in the I-Ching. See below.) Wang Yin-chi % £/\ Z— simply uses the words (particles) yU , , y{leh 0 to render the word yU -+~ without elaborating on its function (1974:13). No matter how scholars interpret the word yU in sentences such as "ft f ^ » it nas one important characteristic that we must not lose sight 9 of, that is, it never functions as a main verb in the Shih Ching. A sentence such as '* (the king goes) ' is unattested in the Shih Ching. In the cases where a locative term follows the word yU -j~ , there is always a verbal element, e.g., jjj^ ^ '-/^ ^\ . When yU appears in sentences such as ^ %j f ff*, sentences to which the gloss j" , ^5 is frequently applied, the expression following yU is invariably verbal.*^ The general rule that yU does not function as a main verb can also be applied to the Ch'un Ch'iu jfti^ and its three commentaries. A preliminary investigation made with HY Index has failed to reveal any case where yU -j~ has to be interpreted as a full verb.*''' In the Kung Yang Chuan ^ ^ flj^ and Ku Liang Chuan |^ ^ ^ there are sentences such as 1 f , £3 ', but yU -p is not a full verb here since this is a commentary on the Ch'un Ch'iu sentence: # £ */£ f% f ^ '(SL 47/& 1/4 £1,4 £) 'a reception house was built for the King's daughter outside (the city wall).' (Legge, p.72) yjl is not found as a full verb in the Shang Shu, I Ching or bronzes either. As in the Shih Ching examples cited above, yjl ^ is occasionally found - 103 -preceding verbal expressions. In the chapter 'Ta Kao' of the Shang Shu o a, we f ind: '....to go forward to restore tranquillity and to perpetuate the plans of my father.' (Legge, p.366-67) '....to achieve the serene (dead) Wu's planned work. (Karlgren, 1950a:37) f "if "> % & ?f f'ty H S (SSTC 270199) 'I will now go forward with you from all the states, and punish those vagabond and transported ministers of Yin.' (Legge, p.367) 'I shall with you various states, go and attack the fugitive and thrown-out (king's servant=) grandee of the Yin (house).' (Karlgren, 1950a:37) (These two examples are cited by Karlgren as evidence in favor of the ' , ^£j<£, ' interpretation (1942:100). Notice in their translations, both Legge and Karlgren interpret vjl as a full verb. Ch'U Wan-li/^,^'^ also renders ^ as 'to go' (1972:72).) 'How should I be all for the oracle of divination and presume not to follow your advice?' (Legge, p.374) 'How should I (first) explore to the utmost the oracle, and then dare not go and follow it?'-'-2 (Karlgren, 1950a:39) In the I-Ching, we find: (HY 23/36/^) 'Darkening of the light during flight He lowers his wings.' (Baynes 1950: Vol.1, P.151) - 104 -9 4j, > ® frfc (HI 23/36/ ) The superior man does not eat for three days on his wanderings. (Baynes 1950: V61.I-, p.151) Ytl is found in the bronzes in the following types of sentences. fit ff rf*i\H&m i wei / king / yU / attack / Ch'u / earl / being at / Yen While the king was on the expedition to attack the Earl of Ch'u, he stationed (lit. being at) Yen.... (San Tai > IK. , ChUan 9, p.26, Ling Kuei /C^ffi ' king / order / Yi Tzu / arrive / west / direction ?» ' • yU / inspect (San Tai • >'fv\i , ChUan 4, p.7, Fa Chiao Ting flfc ft] ifjft The king ordered Yi Tzu to arrive at the west to inspect. l^f& 1 ..ft, i-ssu / prince / order / Hsiao Tzu 2 /go h VA ^ A fit .|13 lead / people / yU / Chin ,fK On i-ssu day, the prince ordered Hsiao Tzu |§[ (?) to go ahead to lead the people (army) to Chin. (San Tai Z- k\, , ChUan 13, p.42, 'I' Mu Hsin Yao '| ^ (The reason for interpreting hsien as 'go ahead' is discussed on p.108). Before proceeding to investigate the linguistic behaviour of y_U ~f in the O.B.I., we may briefly make a conclusion on the presumed verbal function of y_U . YU never functions as a full verb in the early corpus - 105 -we have investigated above. The only verbal function yjl has, if we are justified in terming it 'verbal', is that it can precede the verb (or a verbalized noun) to show that the action of the verb is in process or is the main point of an entire movement. For example, y_U "T in ' J\% ^ ' shows that the military expedition is in process; yU in ' 'f$. — i ' shows that 'to proceed' is the main point of an entire movement. This interpretation of yU throws light on the rationale for Mao's and Cheng's l|p glossing of yU as wang One of the character istics of the word wang is that it frequently appears as the first verb in a series. And, in the classics, only extremely rarely does it take a 14 locative complement although it sometimes functions as a full verb. It should be noted that the word wang , when functioning as the first verb in a series, has a very light information load. A sentence such as ' J ^ ^ ^ ' (the king goes hunting) is cognitively equivalent to ' £ v-j> ' (the king hunts). (In the O.B.I., both ' £ j>% <® ' and ' ^ \J? ' occur quite frequently). There are even sentences such as J {£/ 'the king should go to go out' (ch'u ^ is used in the sense of ch'u tung %%t} (to take the field), e.g., I ^ 1887, I Zi 6530). The fact that wang ^JL , in the above structure, ceases to be a verb carrying a full information load, makes it functionally similar to yjl -f . It may be in this sense that Mao and Cheng jjjip glossed yU ~^ as wang . (Also ^p *VjwaY ^ and ^"J *Yjwang are probably etymologically related. They share a final flexion Cj1^ I^j^^ff^ ^ ' fact was drawn to the author's attention by Takashima in his comments to a draft version of the present thesis.) Our investigation has in fact shown that it is only in this sense that such a gloss is legitimate.''5 - 106 -It may seem that we have gone too far into the investigation of the presumed full verb usage of yjl in the classics and bronze inscriptions. But since all these materials and the O.B.I, are in the same lineage, the usage of yjl in these materials may throw some light on its function in the O.B.I. In any case, the determining factor in the interpretation of yd in sentences such as ' ^,|t ^ ^fij ~f ^|^>' is the linguistic behaviour of yjl in the O.B.I, itself. To the best of the author's knowledge, the first scholar to interpret yu ^ as a full verb 'to go' in the O.B.I, was Yang Shu-ta ~%t) 1ii . He cites three inscriptions as examples where yjl J can be interpreted as wang fa (1974:12). If the officials 'go' to perform a burn-sacrifice to the north tablet (spirit), (they or we) will not encounter heavy rain. If the king walks from J^'j (?) to Ku, he will have no disaster. oifU £ -5- a We should order Ch'ueh to go first to 3(0 (Yang*4^) probably interprets hsien as an adverb, and yjl as the main verb, i.e., 'first go to'. But there is evidence showing that hsien £^ , in the O.B.I., can function as a main verb having the meaning 'to go ahead'. See page 108.) In the first example, vjl ^ precedes a verbal structure as it does in the classics and bronzes. In the second example, the word yU appears - 107 -in a 'V + B *f ' pattern in which both chi |f and yU -f are coverbs (preposition-verb). In the third example, although it is possible to interpret yU -j- as the main verb, another analysis which takes hsien ^ as the main verb seems more plausible. We will return to this point later. Han Yao-lung ^j- ft^. in the article "Chia Ku Pu Tz'u Chung Yd Tzu Yung Fa T'an Chiu' ^ iff f |^ ^ f % cites twelve inscrip tions in support of the theory that yU -j" can function as a verb 'to go' (1973:11). However, among his twelve examples, there is not a single one where yjl has to be interpreted as a main verb. I-**? • ? t * f ^ & T~V m (1) If the officials go to perform a burn-sacrifice to the north tablet (spirit), (they or we) will not encounter heavy rain. P t S f 4 £ f P (2) We should order Ch'Ueh to go ahead to f 4( A* 'h f (3) We should call upon our people to go ahead to (?) (4) We should not call upon our people to go ahead to °& (?) (5) We shouij.g not call upon Lady Hao to go ahead to P'ang to offer people. (.6) We should call upon Lady Hao to go ahead to offer people at P'ang. Zi I f f t H f f i K ^ t' (7) We should not call upon Lady Hao to go ahead to offer people at P'ang. - 108 -I f f /fi e (8) I (the king) should station at P'ang. (9) *The king (should) go to Shang. Or: The king (should do a certain thing) at Shang. I ^ it L f m (10) We should not send people to Ch'a. £ it A. f -it. ^ (11) If the king sends people to Chih, there will be approval. i A. f f (12) We should send people to Hua. In inscriptions no. 8, 10, 11 and 12, there are other full verbs, i.e., B 'to station' and 'f-^L'to dispatch, to send', which can be interpreted as the main verbs with vjl j" as a preposition. Thus, there is no need to interpret vjl -j~ as the main verb. From examples no. 2 to no. 7, the word vjl ~f is invariably preceded by the word hsien %j (Han seems to interpret hsien  /7u as an adverb of the presumed verb vjl •j" ) . One cannot but wonder why the presumed verb yjl -f co-occurs with the presumed adverb hsien 'fcj so frequently (while the verb wang \ j- Is never preceded by hsien ^ in the O.B.I.). The facts that hsien l^_J frequently precedes yU and that it does not precede wang lead the author to suspect that hsien ^ is a full verb, having a meaning 'to go ahead'. Such an interpretation is not without support. ting-yu / crack / Ma / perhaps / go ahead / not / unlucky If our horses (a military unit) go ahead, (they) will not be unlucky. Nan Ming (% eft 682 ' i-ssu / test 7 hui /p (?) / go ahead / statelet / t It is 77 who will go ahead to the statelet. „ ... n /r„ y Nan Ming ffi 9f\ 453 In the above inscriptions, hsien /%_J is obviously a full verb. Thus, the structure hsien yU ^ -J- can be analysed as 'V + preposition' rather than 'adverb + verb'. Inscriptions no. 5, 6 and 7 are, in fact, counter--109 -evidence of Han's theory. In no. 5, the structure is '... ~f /^jf" X^,' while in no. 6 and 7, the structure is '... 'fijffi /v f $ff, '• The fact that the structure ' ~f ' can be placed either before or after ' 'f^ X^' shows that ' "X /ffff, ' would better be analysed as 'preposition + locative complement'. The only inscription not yet discussed is no. 9: ' ;t f f^f '. If we assume that this is a complete sentence and take the surface value of it, yu -j" has to be interpreted as a full verb. Nevertheless, if the practice of deletion is taken into consideration, the validity of interpreting yjl ^~ as the full verb would become questionable. Let us look at some inscriptions cited from the Ping Pien: keng-tzu / crack / Nei / should not / yU / Tzu Ping 122 (2) If we assume that any charging statement has a verb and this verb must be expressed, yd -j~ ought to be interpreted as a verb in the above inscription. However, the tui-chen counterpart of Ping 122 (2) reads, test / perhaps / rest / yU / Tzu We should perhaps rest at Tzu. Ping 122 (1) It shows that the verb in Ping 112 (2) is not yU, instead it is ejfc-'to rest' which is deleted. Examples of the above type can easily be multiplied: v 'ft it yU / Shu At' • Shu Ping 324 (4) t / ytt / Chia Pin / tes  yU  / capture Ping 324 (3) We will capture (deer) at Chia Pin. - 110 -(.On the same plastron, there are divinations about whether there are deer ( l%y ) or not. Presumably 'deer' is the unexpressed object of the verb |t(=4i) 'capture'.) test / should not / yU I Ying / Tz'u (The king) should not (go into) Ying and Tz'u. Ping 521 (14) £4 test / king / enter / at / Ying 7 Tz'u The king should go into Ying and Tz'u.17 Ping 521 (13) (For the interpretation of the graph (word) , see YU Hsing-wu 1979:174-175.) Keeping the practice of abbreviation in mind, whether yjl in the inscription ' £. 'f 1 functionsyas a verb is disputable. Unfortunately, as this inscription appears on a very small piece of plastron (or bone), the tui-chen counterpart and the larger context are not accessible to us. way: Inscriptions similar to ' X- T ffl ' can be interpreted in a similar chia-wu / crack / Cheng / yU / River Ping 203 (16) chia-wu / crack / Cheng / should not / yU / River Ping 203 (17) First of all, we must recognize the fact that the word )&[ 'River' does not only represent a spatial (physical) entity but also a supernatural being, i.e., the River god. (See Chia P'ing ^ -°f .' 1980:210.) In fact, the number of cases in which ho represents a supernatural being is much greater than those in which it represents a spatial entity. Thus, the inscription ' f ' - Ill -is probably an abbreviated form of * -p (to perform an yu-offering to the River god). To assume that ho ' represents the River god rather than a spatial- entity is not mere conjecture. On the same plastron, we find the following inscriptions. jen-yin / crack / Ch'Ueh / test / River(god) / harm / king The River(god)will harm the king. Ping 203 (20) j en-yin / crack / Ch'Ueh / test / River(god)/ not / harm / king The River(god) will not harm the king. Ping 203 (21) Thus, in the inscriptions T ^"J and & "f , the word yU -J" is probably 18 a preposition rather a full verb. Takashima also lists some inscriptions in which yjl is interpreted either as a verb or a particle (1973:156), but he has not elaborated on the distinction. Let us look for the presumed full-verb usage of yU in all these inscriptions: test / king / yjl / Kung." Ping 3 (9) The king should (station) at Kung. should not / yjl / Kung / station Ping 3 (10) The king should not station at Kung. - 112 -test / Yen"7 yU / Mien ytl n Ping 120 CIO) Yen, at Mien, Yen / should not / yU / Mien Ping 120 (11) Yen should not, at Mien, test / Yung / cut grass / yU / Ch'iu Ping 1A1 (13) Yung should cut grass at Ch'iu. M $ %ift ^ £ ^ "ft test / Yung / cut grass / should not / (yU) / Ch'iu Yung should not cut grass at Ch'iu. Ping 141 (14) •M* U V $ f f C| 5*. chia-wu / crack / Cheng / yU / River(god) Ping 203 (16) To the River(god) -ff h ^ kx> f k * f» rh«i chia-wu / crack / Cheng / should not / yU / River(god) To the River(god) / we should not Ping 203 (17) test /..... / woman / yU / Tun T Ping 326 (7) woman at Tun. - - 3 not / vjl / Tun should not at Tun. should  ytt 7 Tun Ping 326 (8) Yung / cut grass / yU / Ku Ping 396 (6) Yung should cut grass at Ku. should not / ytt / Ku Ping 396 (7) (Yung) should not (cut grass) at Ku. 1/ fo • t H test / Kung / cut grass / should not / yjl / ^ (?) Kung should not cut grass at %• C?). Ping 413 (.21) i-ssu / crack / Chung / test / Kung / cut grass / yU / S=|, (?) Kung should cut grass at ^§ (.?). Ping 413 (22) H JU f f '} have / ensue / yU / River(god) Ping 443 (5) have ensuing , to the River(god). y I ff fj b f r r? 1.7 perhaps / ensue / should not / yU / River(god) Ping 443 (6) 19 .... perhaps ensuing should not .... to the River(god). test / should not / yU / P'ang Ping 510 (8) should not at P'ang. test / at / P'ang Ping 510 (7) should at P'ang. (Takashima lists only the negative sentences since his major concern is with the negatives. The positive counterparts are supplied by the author.) In Ping 3 (9) and (10), since the verb jj? 'to station' appears, it does not seem necessary to interpret yjl as a full verb. Moreover, if we look at another tui-chen pair engraved on the same plastron, the motivation for interpreting yU as a preposition becomes stronger: chi-wei / crack / Ch'Ueh / test / we / yU / Chih / go into / station. We should go into Chih to station. Ping 3 (19) test / should not / yU / Chih / station Ping 3 (20) We should not station at Chih. - 114 -The occurrence of the verb A.'to go into' shows that yU •j' functions as a preposition rather than a verb in Ping 3 (19) and (20). Ifitis legitimate to take Ping 3 (19), (20) and Ping 3 (9), (10) as a group, as indeed these inscriptions appear to be, we may supply the verb -s\ 'to go into' to Ping 20 3 (9), (10) and interpret yU as a preposition. Ping 203 (16) (17) ' f \f) ' and ' -f ' have already been discussed on p.110. Similar to Ping 203 (16) (17) are Ping 120 (10) (11) ' £ 1 k$ ', ' i fpj • and Ping. 510 (8) ' f ', 'fe O) $' • i Although there is no clue as to what the verb is, it seems highly probable that some verbs have been deleted?1 Ping 326 (8) ' 4? 1 =jf 'is only an abbreviated form. In its positive counterpart, there is one more graph ( -kr), i.e., 0 • Chang Ping-ch'Uan transcribes as iffl , probably taking it as a negative. This transcription is not acceptable since it violates the general tui-chen pattern of positive versus negative. We don't know exactly what graphs precede the graph >^ . In the Yin HsU  Wen Tzu Cho Ho 211 •^h^l^'i* > we find the following inscriptions: chi-ssu / crack / Cheng / test / fang-sacrifice / woman / at / Tun 22 We should, fang-sacrifice a woman at Tun. test / fang-sacrifice / woman / should not / at / Tun We should not fang-sacrifice a woman at Tun. i _ -k The obliterated graphs in front of M -j" |T in Ping 326 (7) may be ^ Ping 141 (13) (14), Ping 396 (.6) (7) and Ping 413 (21) (22) are the type of inscriptions in question. -115 -This lengthy excursus on the presumed full-verb usage of yU has demonstrated that there are almost no instances in which yU "j" must be interpreted as a full verb 'to go1. In both the three classical texts where it is employed most frequently and in the archaic inscriptions, yjl functions as no more than an auxiliary verb signifying that the verbal element following is the purpose of an entire action. If the above obser vation is correct, sentences such as ' fL'f: % ~f "WL>^ and ' ^Lf| ^ ^ *f should not be analysed as composite sentences since to do so would require us to interpret ' J~ ' and ' /J1J "X ^jg, ' as the main clauses where a full verb should occur. Furthermore, it has been shown above that the negative wu lJTJ in the O.B.I, has a peculiar behaviour. It can be placed in front of the con stituent which is the focal point of negation. That the inscription ' ^£ ^ 17] "J" exemplifies this structure rather than the 'normal' structure ' * l^if 7// 'ff f ^s a Pro°f that the focus of the negation is on the locative ^jp . On Ping 396, there are the following pairs: ^ % ff Yung / cut grass I at I Yuan 396 (1) Yung should cut grass at YUan. %n f% kit ff ft Yung / cut grass /-should not / at /.YUan 396 (2) Yung should not cut grass at YUan. (For the interpretation of the graph(word) ^ , see Yu Hsing-wu 1979:331-333.) test / Yung / cut grass /.at / Ch'iu 396 (3) Yung should cut grass at Ch'iu. Yung / cut grass / should not / at / Ch'iu ; 396 (4) Yung should not cut grass at Ch'iu. - 116 -test / Yung / cut grass /at / f (?) 396 (5) Yung should cut grass at ^ (?)• Yung / cut grass / at / Ku 396 (6) Yung should cut grass at Ku. t; * f •? ^4 should not / at / Ku 396 (7) t (Yung) should not (cut grass) at Ku. This array clearly demonstrates that the locative (the place where grass cutting should be done) is the main concern of these divinations. The occurrence of wu right before yjl •f is a 'marked1 construction used to single out the focal point of negation. In our corpus, this syntactic pattern is quite rare. A more common construction which has the same function is to put the constituent conveying the focal point of negation in front of the verb, frequently in the initial position of a sentence. For example: Ping 3 (9) (10) (19) (20) —- J -*K n 1 the king should station at Kung. ' ^ £f ) jlf ^« (the king) should not station at Kung. ' ~$C h ' We should go into Chih to station. We should not station at Chih. (It is assumed that the term Tfj^ 'we' includes ~£, 'the king' in this context, i.e., the king and his army.) It should be noted that although the negative wu ^ can be placed in front of a. constituent which is the focal point of negation, it is not the case that wu 7^? is always used in this way. O7 - 117 -CHAPTER THREE CONDITIONAL AND SIMULTANEOUS-SUCCESSIVE SENTENCES I (NON-RITUAL-SACRIFICIAL VERBS) In this chapter, we are going to deal with two kinds of composite sentence, namely, conditional and simultaneous-successive sentences. These two kinds of sentence-, however, will not be investigated under separate headings. The reason for this is that: (1) without formal markers, it is sometimes difficult or even impossible to distinguish these two kinds of sentence , and (2) these two kinds of composite sentence are so similar that, in classical Chinese, sometimes the same connective can be used in either kind of sentence. (See p.143). Thus, they are to be discussed together, although efforts will be made to distinguish them through the translations. Furthermore, as logical relationships between the clauses in O.B.I, containing sacrificial verbs are rather obscure andltheiriiriscript'ion requires a different approach, the analysis and discussion of such inscrip tions will be dealt with separately in Chapter Four. As a subcategory of subordinative composite sentences, cause-and-effect sentences have been discussed in Chapter Two. Ther.e^is no-need fotset up a separate section for this kind of sentence in the present chapter. Instead, it will be discussed in connection with the conditional sentence. - 118 -I. A GENERAL DESCRIPTION OF CONDITIONAL AND SIMULTANEOUS-SUCCESSIVE SENTENCES These two types of sentence constitute a large part of the total of composite sentences. They are used to divine about: A. What actions should be taken in a given situation, as in: i .Y'**r» fis. ^ *« test / Chih Chia / open / Pa / king / follow f;X;Chihi Chia'.breaks into (lit. open) the Pa, the king should follow him. Or: Chih Chia broke into the Pa, the king should follow him. Ping 275 (5) hsin-mao / crack / Pin / test / Chih Chia / open / Pa ±* £ * /^ii ^-i fl« king / should not / wei / this / follow (If) Chih Chia breaks into the Pa, it should not be he that the king follows. Or: Chih Chia broke into the Pa, it should not be he that { the king follows. ^ ^ (g) hsin-mao / crack / Pin / test / Chih Chia / open / Pa king / hui / this / follow / five / month (If) Chih Chia breaks into the Pa, it should be he that the ^ king follows. (Divined in the) fifth month. 276 (6) - 119 -Or: Chih Chia broke into the Pa, it should be he that the king follows. (Divined in the) fifth month. 1K W* ^ 4? *jr ting-ssu / crack / Cheng / sick / foot / exorcise to / Father Keng Ping 541 (3) (So and so) is suffering from an ailing foot, we should . perform an exorcism to Father Keng. ft* ^ it ^ ^ & JL %f f sick / foot / should not / specifically 7 exorcise / to Father HsinT Ping 541 (4) (So and/so) is suffering from an ailing foot, we should not specifically perform an exorcism to Father Hsin. (The reason for excluding the conditional interpretation of Ping 541 (3), (4), i.e., 'if so and so suffers from an ailing foot, ...', is discussed on p.311, fn.13) B. A consequence of a certain activity or event, as in: chia-ch'en / crack / Ch'Ueh / test 7 next / i-ssu say / Prince (?) / driveV arrive / at / ting-wei / harm Ping 302 (.13) On next i-ssu day, (if) we call upon Prince ^ (?) to drive 1 (the enemies), arriving at the ting-wei day, we will harm (them). - 120 -< nhft V* Us Sf*. f* next / i-yu / yu-offer / behead / from / Hsien / approve Ping 41 (18) On next i-yu day, if we yu-offer beheaded victims (to ancestors) starting from Hsien, (they) will be pleased. Ttf U 1> %t ft&^M t*U fee wu-wu / crack / Chung / test / P'an / go / come / have no A fa misfortune Ping 130 (1) During P'an's travelling to and fro, there will be no misfortune. test / P'an / go / come /.perhaps / have / misfortune Ping 130 (2) During P'an's travelling to and fro, there will perhaps be ^ misfortune. keng-tzu / crack / Ch'Ueh / Lady Kuo / bear / good Ping 96 (16) Lady Kuo will give birth to a child, it will be good. Lady Kuo / bear / not / perhaps / good Ping 96 (17) Lady Kuo will give birth to a child, it will perhaps not be good. V Sentences of this kind, incorporating the consequence of an activity or event, can be further sub-divided into two categories: - 121 -(.1) The activity stated in the subordinate clause is intended to bring about the consequence/effeet found in the main clause, for example, Ping 302 (.13) ' Z» £ €? 4" j| , J -f T ^ ' (see also the inscriptions cited in footnote no'.l). There is a quasi-'cause and effect' relationship between the subordinate clause and the main clause. (2) The activity stated in the subordinate clause is not intended to bring about the consequence represented by the main clause, as in Ping 130 (2) ' -|}$_ %t^, jt f 4$) '• Obviously, 'misfortune' is not something the Shang people intend to bring about during P'an's travelling to and fro . 'There will perhaps be misfortune' simply represents a situation, rather than an 'volitional effect', of the activity 'travelling to and fro'. There may be more than one main clause in a conditional or simul taneous-successive sentence: i-ch'ou / crack / Ch'Ueh / test / chia-tzu / cut into / i-ch'ou king / dream / shepherd / stone / mi-deer / not / wei / misfortune wei / blessing Ping 96 (1) During the transitional period between the chia-tzu and i-ch'ou day (lit. the period when the chia-tzu day cuts into the i-ch'ou day) the king dreamt of shepherding stone mi-deer; it does not (mean) misfortune (but) blessings. (For the interpretation of the graph(word) , see Takashima 1979:54, fn.l9a.) - 122 -t "it. ft 8 test / king / cut(?) / three / Ch'iang / at / tsu-stand not / obstruct / please Ping 96 (11) (If) the king cut(?)-sacrifices three Ch'iang-tribesmen on a tsu-stand, (the spirits) will not obstruct (but) approve (the 2 king's activities). In the majority of conditional or simultaneous-successive sentences collected in the Ping Pien, positive-negative polarity is expressed in the apodosis. f W fi *Fj. 2* £* F-S/^^/ • test / Chia / go / come / perhaps / have / misfortune During Chia's travelling to and fro, he will perhaps have misfortunes. Ping 32 (26) < kuei-ch'ou / crack / Cheng / test misfortune / king / prognosticate / say / no / misfortune / Chia / go / come / have no Ping 32 (27) During Chia's travelling to and fro, he will have no misfortunes. The king prognosticated and said, "there will be no misfortunes." chia-wu / crack / Ch'Ueh / test / king / present / this jade (bundle) / Hsien / obstruct If the king presents this jade (bundle), Hsien will obstruct (his activities). Ping 139 (1) - 123 -j chia-wu / crack / Ch'ueh / test / king / present / this / jade (bundle) / Hsien / not / obstruct If the king presents this jade (bundle), Hsien will not s obstruct (his activities). Ping 139 (2) However, there are also some tui-chen pairs in which the positive-gative polarity is expressed in the protasis. ' y:7k '< * U Fl* H* £» 4 » keng-wu / crack / Nei / test / king / should not / make s&, r* <u n ft settlement / being at / here / ti-god / approve Ping 93 (4) If the king does not make a settlement here, ti-god will approve (or, will be pleased). Or: Let the king not make a settlement here and then ti-god 3 will approve (or, will be pleased). fy/kU U ii IN* I g. keng-wu / crack / Nei / test / king / make / settlement / ti-god approve / eight month If the king makes a settlement here, ti-god will approve (or, will be pleased). Ping 93 (5) n% i. li tf fi test / king / make / settlement / ti-god / approve If the king makes a settlement, ti-god will approve (or, will be pleased). Ping 93 (6) - 124 -\ test / ... / should not / make / settlement / ti-god / approve Ping 93 (7) If (the king) does not make a settlement, ti-god will approve (or, will be pleased.) Or: Let the king not make a settlement and then ti-god will approve (or, will be pleased.) jen-tzu / crack / Cheng / test / we / perhaps / make / settlement ti-god / not / obstruct / approve Ping 147 (1) If we will perhaps make a settlement, ti-god will not obstruct but approve (or, will be pleased). kuei-ch'ou / crack / Cheng / test / should not / make / settlement ti-god / approve Ping 147 (2) If the king does not make a settlement, ti-god will approve (o.r, will be pleased). Or: Let the king not make a settlement, and then ti-god will approve (or,i will be pleased). kuei-ch'ou / crack I ... I test / we / make / settlement / ti-god not / obstruct / approve If we make a settlement, ti-god will not obstruct but approve (or, will be pleased). Ping 199 (11) - 125 -\ kuei-ch'ou / crack I ... I test / should not / make settlement I ... I approve Ping 199 (12) If we do not make a settlement, (.ti-god) will approve (or, will be pleased). Or: Let us not make a settlement, and then (ti-god) will approve ^ (or, will be pleased). These inscriptions obviously cannot be interpreted as 'cause and effect' sentences since *'we/the king should not build a settlement, because ti-god will be pleased' does not make sense at all. On p.82 & 84 "of Chapter Two, we have already discussed, by citing Serruys' and Takashima's studies, the conditional usage of the prohibitive negative wu . Their theory is certainly applicable to the inscriptions in question. There appears to be a difference between this type of conditional sentence and ordinary conditional sentences. In a normal conditional sentence of which the protasis represents a controllable activity, the Shang king (or diviner) merely proposes a possible choice without strongly or explicitly expressing his attitude towards the activity. However, in the type of conditional sentence exemplified above whose protases represent controllable activities, the Shang king's (or the diviner's) attitude towards the activities is strongly and explicitly expressed by use of the modal ?D , e.g., we should not make a settlement. All these four tui-chen pairs divine about the building of a new settlement. Whether all these inscriptions deal with the same proposed settlement is something we cannot be sure of. The diviner of Ping 93 (4) (5) is Nei while the diviner of - 126 -Ping (1) (2) is Cheng ^7 and furthermore, the divinations were made on different days, the keng-wu (.seventh day in the kan-chih calendar), the jen- tzu (forty-ninth day in the kan-chih calendar) and the kuei-ch'ou (fiftieth day of the kan-chih calendar) respectively. Interestingly enough, upon checking the inscriptions concerning the building of settlements ( A rt ) , it is discovered that none of these inscriptions takes the common pattern in which negation is expressed in the apodosis, i.e., '* £•/ ^ ^ ^ "2^2. (*if the king/we make a 4 settlement, ti-god will not be pleased'. Unfortunately, due to the paucity of this type of sentence, it is difficult to provide an explanation. - 127 -II. THE CONTRAST BETWEEN AN INTENDED EFFECT AND AN UNCONTROLLABLE CONTINGENCY In many cases, it is easy to determine whether or not the result expressed in the main clause is something the Shang bring about intentionally by performing a certain activity. Typical examples are 1 1 (to attack) and its intended result (harm); or 1 ^T3E- ^ ' (to travel to and fro) and the undesirable condition ' ^ l$p) ' (have misfortune):! cited above. However, there are also some cases where both interpretations seem to be possible. Sentences where (^T) 'to rain' appears as the main clause sometimes pose problems. Depending on the particular situation, rain may be something either desirable or undesirable, i.e., the intended effect of an action or an unwelcome possibility. If 1* H 6 A « I * rf $<=*]) chia-yin / test / yu-offer / ascend-sacrifice / sui-cut-sacrifice i|_ '.M tag T'ai I / encounter / rain During the yu-offering, ascend-sacrifice and sui-cut-sacrifice to T'ai I, we will encounter rain. Or: If we perform a yu-offering, ascend-sacrifice and sui-cut-sacrifice to T'ai I, then we will encounter rain. .j, Ts'ui 794 U< H & A *• 4 M T * 0u) <f kuei-hai / test / yu-offer / ascend-sacrifice / to / (Shang) Chia encounter / ram (During) the y_u-off ering and ascend-sacrif ice to (Shang) Chai, - 128 -(we) will encounter rain. Or: If we perform a yu-offering and ascend-sacrifice to (Shang) Chia, then we will encounter rain. , . ••a? Chia N-P 750 At first sight, both interpretations seems equally possible, but as will be pointed out later, the alternate interpretations posed here will not be the correct choices. In order to choose between them, we have to study the common expression There are two characteristics of this expression. (1) ' ^) ' sometimes co-occurs with the word ch'i which is frequently, though not necessarily, associated with undesirability, e.g., Chien-^ 17.10, Ts'ui 721 and Ch'en Pjfc. 4. (This characteristic feature of j% is observed by Serruys (1974:25).) While holding an opinion different from that of Serruys* on the interpretation of ch'i, the author agrees that ch'i, in most cases, can be used as a criterion to determine whether a situation is considered undesirable by the Shang providing there is no counter-evidence. (For a discussion of ch'i, see p.142-) But the expression ^NjJ| $S) is never found in collocation with . On the contrary, c^T) sometimes co-occurs with chi (auspicious), e.g., I fyf^ 48, Ho 31 and IHayashi' "fcjv 2.16.22, while rfg never does. Taking these two observations together, we can suggest that |S?represents something undesirable. (2) In the Shang and Chou times, dancing is a well-established practice for invoking rain. - 129 -. .1 f a °A«i ft m « test / next / ting-mao / present / dance / have / rain I ZJ 7233 On the next ting-mao day, if we present a dance, we will have rain. 7 - * call upon / dance / have / rain Chin 638 (If) we call upon (so and so) to dance, we will have rain. To teach the huang-dance and lead (the people) to dance (because of) the drought. (SSCCS, Chou Li ffl \\j , chuan 12, p.22) 'Raining' is definitely the intended result of the dance. However, sentences incorporating both '* ' and the expression under discussion ' CK) / ' are n°t attested in the O.B.I. Similarly, the ritual .jjc (to invoke), is frequently used to invoke rain directly or indirectly. keng-wu / crack / test / to / great / tablet / invoke / grain rain Jin 2366 (If) we invoke for a (good) grain harvest to the (spirits of the) great tablets, it will rain. invoke / to / River (god) / harvest / have / rain Ts 'ui "^f* 834 (If) we invoke for a (good) harvest to the River (god), we will have rain. (See also Nan Ming 424, 426; Ts'ui ^ 121; and Chia ^ 1275. List inexhaustive.) - 130 -perhaps / invoke / (?) / gate / have / big & fa rain Chia f 1259 (If) we perhaps invoke for rain at the X. (?) gate, we will have heavy rain. However, the expression 3*i[5_.ft& never co-occurs with this verb either. That the expression jjljf fifcj never co-occurs with the verbs wu J|p and ch'iu $L of which 'raining' is an intended result suggests that jfl f^? is something unintended. (The direct object of , i.e., the thing prayed for, is not restricted to vjl jfcj . However, the direct or indirect soliciting of rain seems to be the most common purpose of this ritual.) To sum up, the expression jjjf pp7 does not represent the intended effect of an action and its co-occurrence with ch'i jjL suggests that it is possibly something undersirable. Such a generalization is confirmed by the following inscriptions. pmg-yin / crack / i (?) / test / next ting-mao / king / perhaps / teach / not / encounter rain Pu f 501 On the next ting-mao day, the king will perhaps teach I (the mass); he will not encounter rain.5 - 131 -king / perhaps / hunt / not / excounter / rain The king will perhaps hunt; he will not encounter rain. Ching Jj. 3851 / ~^JC_ 'Teaching (the people martial activities)' and ^ 'hunting' are not generally considered rain-soliciting activities. Quite the contrary; 'raining' poses an obstacle to such activities. It is more than natural to interpret /^jjl^ as a desirable but uncontrollable possibility, rather than an intended consequence, of and 09.. Once this idiomatic usage of the expression ij| (jt^ / is made clear, we can reject the alternate interpretations of Ts'ui 794 and Chia 750 on page 127. In contrast to , the expression ' ~% $0 ' represents the hoped-for result of rain-soliciting ritual-sacrifices. at stake / Tsai / hav 7t f H$i ^n*l* ^ .. , test / this / ping-hstl / burn at stak  have ensue / rain Ping 157 (1) On the ping-hsti day, if we burn at the stake a woman of Tsai, we will have ensuing rain. A bi ^* # A* rri & test / Tsai / have no / perhaps / ensue / rain Ping 157 (2) (If we burn at the stake) a woman of Tsai, we will perhaps have no ensuing rain. < - 132 -test / ch'tl-sacrif ice / Mountain(god)" / have / rain If we perform a ch'tl-sacrif ice to the Mountain(god), we will perhaps have rain. Ping 535 (2) ch'U-sacrifice / have no / perhaps / rain If we perform a ch'U-sacrifice (to the Mountain(god)), we will perhaps have no rain. Ping 535 (3) i% ^« ti7* test / burn at stake / have / rain Ping 469 (1) We should burn at stake (X), for we will have rain. should not / burn at stake / have no / perhaps / rain Ping 469 (2) We should not burn at stake (X), (for) we will perhaps have no rain. test / next / ting-mao / present / dance / have / rain Ping 223 (5) On next ting-mao day, we should present a dance, (for) we will have rain. next / ting-mao / should not / have no / perhaps / rain Ping 223 (6) On next ting-mao day, we should not (present a dance), (for) we will perhaps have no rain. - 133 test / dance / Mountain(god) / have / rain If we (present) a dance to the Mountain(god), we will ^ have rain. Ping 199 (9) test / Mountain(god) / have no / perhaps / rain (If we present a dance to) the Mountain(god), we will perhaps have no rain. Ping 199 (10) test / chao / River (god) / burn-sacrifice / to / $(!£ (?) ft ^ ti have / rain Ping 431 (6) If we chao-sacr if ice(?) to the River(god) and burn-sacrifice to (?), we will have rain.*' call upon I ... I wang / to / River(god) / have / ensure /rrain Ping 463 (2) (If) we call upon (so and so) to wang(?) to the River(god), we will have ensuing rain.'' have no / ... / ensue / ... Ping 463 (3) v(We) will (perhaps) have no (rain). . test / we / dance / rain Ping 71 (6) g If we dance, it will rain. This is an exhaustive list of the inscriptions concerning rain-soliciting in our corpus. With the exception of the last one Ping 71 (6), the word yu. or (wang ) is employed in the main clauses. - 134 -Interestingly enough, in all the inscriptions which divine only about the possibility of rain at a future time, (i.e., without mention of rain-soliciting-ritual-sacrifice) the word yu ^jj (or wang t- ) is extremely 9 rarely used. For example, 9 1" m this / night / rain Ping 3 (15) Tonight it will rain. this / night / not / perhaps Ping 3 (16) Tonight, it will not perhaps (rain). < ting-ssu / crack / Hsllan / test / from / this / arrive ? * ^ & ^ * ^ 4 rr,ft at / keng-shen / perhaps / rain Between today and keng-shen, it will perhaps rain. Ping 151 (1) test / from / this / ting-ssu / arrive / at / keng-shen not 7 rain ' Ping 151 (2) Between the present ting-ssu and keng-shen, it will not rain. from / this / keng-tzu / arrive / to / chia-ch'en / ti-god order / rain Ping 381 (7) Between the present keng-tzu and chia-ch'en, ti-god will order it to rain. arrive / chia-ch'en / ti-god / not / perhaps / order rain Ping 381 (8) VBefore chia-ch'en day, ti-god will not order it to rain. It is worth asking why the word yu and its negative counterpart wang c- rarely appear in simple divinations about the possibility of rain at a future time. And as mentioned above, in clauses expressing the intended result of rain-soliciting-ritual-sacrifices, we frequently find the word yu ^ , e.g. *f ($) / % M . Among the various functions of the word yu ( ^ ) , one is a pre-verbal usage which can be rendered into English as 'there is the occasion'. Such an interpretation has strong support from the classics. 6. & H * % *l & 1 t (SSTC 240617) There should be no such thing as a minister conferring favours, displaying the terrors of justice, or receiving the revenues of the country. (Legge, p.334) Yu means 'occasionally there is and occasionally there is not'. (This is a comment on the Ch' un Ch' iu ^ f A sentence ' ijfk ^ (HY 64/|)£18/3 &3C) In autumn, there were yih'. (Legge, p.97).) The further observation that *YjwaJT and ^ *gwak/*yiwak are closely related phonologically leads us to test this interpretation on pre-verbal occurrences of in the O.B.I. - 136 -N ft i* 4 A £ ft £«. test / Ch'Uan / chase / Hstlan / have / catch up If Ch'Uan chases the HsUan (enemies), there is the occasion of catching up. Or: Ch'Uan is chasing the HsUan (enemies), there is the occasion of catching up. Ping 261 (13) •'a test / perhaps / have / come Ping 296 (5) There is the occasion that (so and so) will come. (See Takashima 1973:49-50 for a discussion of the contrast between ( ^ ) & / * & and & I £ . Takashima' interpretation is adopted by the author.) The interpretation 'there is the occasion' for the pre-verbal y_u ^ makes good sense in these inscriptions. However, in the case of yu (tsung) yU 7^. ( ) (-$£7 > the picture is different. It is difficult to conceive why the sense 'there is the occasion' is so frequently used in the inscriptions where yu (tsung) yU 7^ ( ii ) K|T) represents an intended result or desirable situation, but not in the inscriptions which merely divine about the possibility of raining in a future time. Instead of interpreting yjj. ^ as an existential verb, the author proposes to interpret yji as a transitive verb 'to have'. In some cases, we may even interpret y_u ^ as 'to obtain, to possess'. Cf. "J Jjj' -jj^ ^ -f (Mencius HY 10/2A/1) 'Wu-ting had all the princes coming to his court, and possessed the kingdom....' (Legge, p.182). The performance of a rain-soliciting-ritual-sacrifice is for the Shang a means of obtaining rain. Thus, when preparing for such ritual-- 137 -sacrifices, the Shang would naturally divine about whether they can get the rain or not. To take yu as a transitive verb in the expression yutso huo seems the' best interpretation. ' keng-hstl / crack / Nei / test / king / enter / to / Shang have no / make / misfortune Ping 165 (16) Tentative interpretation: *When the king enters Shang, there will be no occasions to cause misfortunes. I ii A^ff ^ y$ ^ test / king / enter / to / Shang / perhaps / have / make misfortune Ping 165 (17) Tentative interpretation: *When the king enters Shang, there will be occasions to cause misfortunes. The expression tso huo ft \w also occurs without ju : hsin-wei / crack / Pin / test / perhaps / make / misfortune J1 **-K **i thirteen month Hou Hsia 73? 7 7.12 (So and so) will perhaps cause (lit. make) misfortunes. (Dividend) at the thirteenth month. ti-god / perhaps / make / king / misfortune Ti-god will perhaps cause misfortunes upon the king. I ZJ 4861 - 138 -In sentences such as /fj? j£- h |* J 71*9 , the structure is fairly clear, we can state with certitude that jti_ ^ is the agent who causes the king misfortune, I.e., agent + (jj^ ) + V + beneficiary + patient. In the case of 1 £ A_ ^ f%\ , j£ ^ ^j*-^^', if we adopt the tentative analysis, the word jj_ 'king' would be the agent while the word yu Tjcj • » though it can be interpreted as the main verb in the surface structure, seems to be devoid of its verbal function; in a semantic analysis, the main verb is tso 'l^t- . Though both syntactically and contextually possible, this analysis fails to account for the presence or absence of the word vju . An investigation of the inscription incorporating the expression tso huo 7 has revealed the following phenomenon. Whenever the subject is a supernatural being, the word yu. ^| does not appear; whenever the subject is not a supernatural being, the word yu 7^ is employed. For example, o r fa >*> Ml & h (River) HsUan / not / make / this / settlement misfortune HsU 1^ 4.28.' The (god of) the River HsUan will not cause (lit. make) misfortune upon this1 settlement. HsUan / perhaps / make / this / settlement misfortune HsU >^ 4.28.4 The (god of) the HsUan will perhaps cause (lit. make) V misfortune upon this settlement. (See also To^^2.476.) - 139 -i-mao / crack / Nei / test / SsuC?) / make / king misfortune I_ 2. 2465 The (spirits of) Szu(?) will cause (lit. make) misfortunes upon the king.1^ In the above inscriptions, the word v_u Fj does not appear. "f1* 5* U ft ft H|ffv V* ^(4 keng-shen / crack / Ch'Ueh / test / we / have / make misfortune I_ C* 6263 We will have misfortunes caused (by X). Or: *There is the occasion that we cause misfortunes. u WsHs 4 i^hai / crack / Ch'Ueh / test / Ch'Ueh / have / make £f US misfortune I_ ZJ 5349 Ch'Ueh will have misfortunes caused (by X). Or: *There is the occasion that Ch'Ueh causes misfortunes. In the above inscriptions where the subjects are not supernatural beings, the word yu ^ appears. In the case of tso , why the selection of the surface subject has influence on the presence or absence of the word yu ^ is a question difficult to answer, since tso'tj\. is a word of multiple usage, and is thus open to different interpretations. This question can perhaps be clarified if we pick a word which allows only a rather rigid interpretation, e.g. , chiang jJJp 'to descend, to send down': - 14G -ill ft, Hi test / dismember (?) I ... I ti-god / not / perhaps send down / misfortune I_ \f\ 36 ' Dismember (?) ..., ti-god will perhaps not send down misfortunes. 6 is ^ ft * test / this / settlement / perhaps / have / send down misfortune Ho A- 312 — rz-This settlement will perhaps have a sending-down of misfortunes. A comparison of these two inscriptions shows that though both tjL 'ip and JL £ are surface 'subjects' in the respective sentences, there is a distinct difference between them. That is, in the first inscription j^L is the agent of the verb chiang while JL in the second is the beneficiary of the verb yu 7^ . In Ho 312, the word yu Tfj is required since a sentence such as * Jr?, ?"zp^ST}is nonsensical. A translation which, more precisely, though not idiomatically, reflects the structure of this sentence is 'this settlement will perhaps have sent-down-misfortunes.' By applying the same analysis to the sentences having the structure: 'non-supernatural-being S + "fa + 1*Ji + ' , we may contend that the expression tso huo h ft ^j^is not a 'verb + object' structure which can function as a predicate; instead, it is a 'modifier + modified' structure, i.e. a nominal, which requires the verb yu with which to form a predicate. The non-supernatural surface subjects (including the locative term jtf, ) are the beneficiaries rather than agents of the predicate 'fyifi'Tri The function and meaning of yu Tj^ in these inscriptions are more 'concrete' - 141 -(i.e. to have, to get) than that reflected in the translation 'there is the occasion'. Another possibility is to analyse the structure \$) I ^ as a clause-object with an indefinite subject, ive. + Indefinite S + "f^. I ^f1 > '.... will have (someone's sending down misfortunes)'. (This alternative is proposed by Pulleyblank in his comments to the draft of the present thesis.) To sum up, although the expressions ' C?9 ', ' ^ (fi) ' and ' ij| f% all refer to the possibility of raining, there are some semantic factors which determine the selection among them./j| c^'meets-.with'ir.ain' "(represents a undesirable contingency (cf. Japanese 'helpless' passive, e.g. ame ni  hurareru) which would interfere with a planned activity; 41 'have/obtain rain' represents the intended result of certain rain-soliciting-ritual-sacrifices while 'it will rain' represents the neutral possibility of rain at a future time.1''' There are some exceptions to this generalization, in particular in the case of YsT9 '> the most unmarked form, which can also be used in situations where either ' ^9 ' or ' ^ ' ^s covmon^-y used. But the above rule holds true in the majority of cases. - 142 -III. A PSEUDO-MARKER OF THE CONDITIONAL CLAUSE CH'I j£ 1. The Sense of Uncertainty Conveyed by Ch'i ^ It can be easily observed that many sentences of the conditional/ simultaneous-successive category are open either to the conditional or simultaneous-successive interpretion. This is particularly true when the subordinate clause represents a controllable activity. For more examples, f* It $h 4>!* If A call upon / many / horse / chasey game / capture (If) we call upon many horses (military units) to chase game, we will capture (them). Or: We/ called upon / are calling upon many horses (military units) to chase game, we. will capture them. Ping 83 (13) tf lit, 1 y V? H|*T* tt « chia-ch'en / crack / Cheng / test / we / attack • Ma Statelet / ti-god / give / we / assistance (If)we attack the Ma Statelet, ti-god will give us assistance. Or: We are attacking the Ma Statelet, ti-god will give us assistance. Ping 114 (7) In general, the protasis (subordinate clause) in a conditional sentence represents actions or events that have not taken place while the subordinate clause of a simultaneous-successive sentence represents an -. 143 -action or an event that is taking or has taken place. However, the distinction is so subtle that in many languages, the same marker may be used in either kind of sentence. A person planning a trip to China can legitimately say, 'when we go to Peking, we should ...', though 'going to Peking' is something hypothetical — an action that has not taken place. The French word 'quand' can sometimes be translated as 'if and sometimes as 'when'. The German word 'wenn' has the function of representing the sense of 'if or 'when'. (Both words are adduced by Chou Fa-kao to illustrate the difficulties in distinguishing conditional sentences from simultaneous-successive sentences (1961:307).) While it is difficult to distinguish these two types of sentences in most cases, the occurrence of ch'i ^~ in a clause seems to be a criterion of the conditional. However, this does not mean that ch'i can only be interpreted as a 'pure' grammatical markers of conditionals. It is equally possible that ch'i is intended to convey a sense of uncertainty, thus marking the subordinate clause as hypothetical. At this juncture, we must examine the interpretation of Serruys who has observed that 'the presence of ch'i marks the proposition or the alternative among possible courses of action which is considered less desirable, less preferred, often positively feared and resorted to only if really unavoidable;' (.1974:25) and has contended 'the translation "perhaps, may be" is not only inadequate, but misleading' (ibid. 58). - 144 -The observation that ch' i is frequently associated with an undesirable alternative is of great value. Yet to reject the widely accepted inter pretation of ch'i as conveying a sense of uncertainty seems to cause some problems. The fact that ch'i frequently collocates with the undesirable alternative does not necessarily mean that this modal itself conveys a sense of undesirability. For the following three reasons, the author argues that ch'i does not convey a sense of undesirability but rather of uncertainty: (1) The word ch'i may co-occur with the words chi (auspicious) and chia \}t) ( , good): pi'ng-ch'en / crack / _Pin / test / exorcise / king prognosticate / say / auspicious / perhaps / exorcise Pu I- 63 (We) should perform an exorcism. The king prognosticated and said, 'it is auspicious, we should (perhaps) perform an exorcism. Or: '... let us perform an exorcism.' Cf. Pulleyblank has pointed out in one of his seminars that ch'i may have the function of softening a command into a wish or exhortation. king / prognosticate / say / god / wei / this / two / month order / sleet (?) / perhaps / wei / ping / not / order / snow wei / keng / perhaps / auspicious Ping 66 (5) - 145 -The king prognosticated and said 'god will order it to sleet (hail(?)) this second month. (If) it is on a ping day that (god) does not order it to snow, it will be on a keng day. It is perhaps auspicious. " wu-wu / crack / .Pin / test / call upon / take / ox hundred / bring / king / prognosticate / say / auspicious bring / perhaps 7 arrive Ping 399 (1) (If) we call upon (so and so) to take ox in the number of one hundred, (so and so) will (successfully) bring (them) in. The king prognosticated and said, 'it is auspicious, (so and so) will (successfully) bring (them) in (and so and so) will p erhap s arr ive. chia-shen / crack / Ch'Ueh / test / Lady Hao / bear / good -Zt j fit fS t=J O W $. ^ A a T ^ kft king / prognosticate / say / perhaps"/ wei / ting / bear tUi Ui ^4 f"*^U £ ,1 £ * good / perhaps / wei / keng / bear / protract / auspicious - s-• 6 *, B ., TV*??* ^* three / ten-day-week / and / one / day / chia / bear / not good / wei / girl Ping 247 (.1) Lady Hao will give birth to a child, it will be good. The king prognosticated and said, '(if) it is on a ting day that she gives birth to a child, it will be good; (if) it is on a keng -146 -day that she gives birth to a child, it will be protractively auspicious. (After) thirty one days, on a chia day, (Lady Hao) gave birth to a child. It was not auspicious. (The child) was a girl. (For the interpretation of ^ ' *jj , prolong', see Y{1 Hao-liang f gL^ 1977.) It is hard to see that the word ch'i in these inscriptions conveys a sense of undesirability. If it does not, can this ch'i be interpreted as a subordination marker? We will return to this question later (see p.159-170). All these inscriptions are prognostications. One may wonder whether the sense of uncertainty is incompatible with the interpretation of the cracks. While it is difficult to give a definite answer since we do not have sufficient knowledge of the Shang divination practice, it seems reasonable to assume that the Shang people were aware of the uncertainty inherent in prognostication. Even the Shang king, presumably the most expert interpretator of the cracks, was surely often proved wrong by subsequent events. Thus, the presence of ch'i in the prognostications may be interpreted as an admission of the residual uncertainty even with the best divination. In the Tso Chuan, there are many cases illustrating the uncertainty of diviners' interpretations of the oracles. - 147 -^ ?fc i. 3-, 1 4- : r < ri fa / & 11 JI fa *•. ji i M © »*» T * 4i (HY 304/^ 25/2 7^ ) Woo-tsze consulted the milfoil about it, and got the diagram K'wan, which then became the diagram Ta-kwo; which the diviners all said was fortunate, He showed it to Ch'in Wah-tsze, but he said, "The (symbol for) a man (in K'wan) is displaced by that for wind (in Ta-kwo). Wind overthrows things. The woman ought not to be married. And moreover, (upon K'wan) it is said, 'Distressed by rocks, holding to brambles; he enters his palace and does not see his wife. It is evil...'. 'Distressed by rocks;' — in vain does one attempt to go forward. 'Holding by brambles; '— that it which trust is placed wounds. 'He enters his palace and does not see his wife; it is evil: '— there is nowhere to turn to." ,T ri, \ (Legge, p.514) (Notice the discrepancy, and thus the uncertainty, among the prognostications proposed by the diviners.) Jit & ^ % 0 % ^ *" * * (SI,67/^ 22/3 £ ) The marquis made him consult it by the milfoil on the future of the boy, when he found the diagram Kwan, and then by the change of mani pulation, the diagram P'ei. "Here," he said, "is the deliverance}" — 'We behold the light - 148 -of the State. This is auspicious for one to be the king's guest. ... Shall this boy in his generation possess the State of Ch'in? Or if he do not possess this State, does it mean that he shall possess another? Or is the thing fore told not of his own person, but of his descendants? The light is far off, and its brightness appears reflected from something else ...." (Compare the repeated employment of the word ch'i Jfc in this citation and Ping 247 (1). Pulleyblank points out in his comments to the draft of this thesis that this passage seems to show that the divination process is difficult and therefore (presumably) subject to uncertainty.)' It is not unreasonable to assume that this sense of uncertainty, expressed so clearly in the Tso Chuan passages was felt by the Shang in their prognostications and that ch'i expresses this sense. (2) On page 35 of the article 'The Language of the Shang Oracle Inscriptions', Serruys lists nine groups of divinatory sentences where ch'i appears in both positive and negative members. (The author would exclude the ninth group from the list since these particular inscriptions do not appear to be tui-chen sentences.) As Serruys admits, this kind of tui-chen pair does nofsupport the theory that ch'i marks the less desirable alternative. For this phenomenon. Serruys gives the following explanation: 'since the majority of divinations are couched in terms of what the diviner already considers desirable, it might be proposed that this exceptional pattern here could be used when such a decision or opinion concerning preference (Legge, p.103) - 149 -or desirability was not expressed or could not be formed. The underlying main clause might then have been: "we shall prepare for the eventuality that'... we shall expect that ..." (.1974:36) Whether 1 a decision or opinion concerning preference or desirability was not expressed or could not be formed' is a question very difficult to answer definitely. In terms of common sense, it seems that the positive outcome of the following two pairs (Serruys' no. 1 and no. 2) is the more desirable: "Lady Ching (her) millet will be harvested. Lady Ching's millet will not be harvested". "(One) might have occasion that they will bring (ivory) teeth. (One) might not have occasion that they will bring ivory". I-While it is difficult to account for this phenomenon in terms of the theory of undesirability, among the eight groups there is a common point which may throw some light on the underlying rationale for the employment of ch'i in both positive and negative sentences. 'The Fang (tribe) might destroy our envoy. The Fang might not destroy our envoy. - 150 -Our envoy might not destroy the Fang. Our envoy might destroy the Fang. (*The graph does not appear in this inscription.) (A) A: 9 Jt ^ ^ 9 ^ A ^ >^ e> ^ cfe JJt] Z ^ 4: ^ "Today it might rain. Today it might not rain. Next Yi-ssu day it might rain. Next Yi-ssu day it might not rain. Next Ting-wei day it might rain. Next Ting-wei day it might not rain." "X (probably Ch'U) leading 30 men from the Ma(tribe) really might shackle (take prisoners) the Ch'iang. Ch'tt's 30 Ma(tribesmen) might not shackle the Ch'iang." - 151 -"Fou might come to visit the king. The Fou might not come to visit the king." -O) J% & 4% * $ |j[ % * * * * £ $ "As to Chih, he might attack You-ku-lo. As to Chih he might not attack You-ku-lo." «8) 4;. a 5 & "The king today will go out. It will rain. It might not rain. 12 It might rain and it might not rain." (All translations are Serruys'.) The first observation we can make here is that all the verbs in these eight groups of sentences take the *p- type negative particles. The apparent exception *mjwang is a negative verb, the contrary of . As Takashima has conclusively demonstrated, the characteristic of the *p- type negatives is that they go with uncontrollable verbs (1973:263). The event ,<<^ 'to rain' is of course uncontrollable. - 152 -As for the verbs (to come, to bring in), ^* (to shackle) and (to attack), although their semantic character may be thought of as 'controllable', there is every reason to suggest that these verbs represent events or actions that were not within the manageability of the king (or the diviner) at the time when the divinations were performed. None of the agents of these verbs, when they are expressed, is the king. The verbs h\t (to attack) and ^ (to shackle) are military activities or events carried out by persons other than the king. The agent of (to come) in group six is Fou ^ , an enemy tribe of the Shang in the first period. We can reasonably assume that these people are not under the direct control of the Shang king (or the diviner) and that these divinations are performed to foretell what these people are doing and the consequences of their activities. The sentences and ~%- ^\ ^f cannot mean 'we should have brought in (ivory) teeth' and 'we should not have brought in (ivory) teeth', for they would then not make much sense when taken as a tui-chen pair. The agent of 'bringing in (ivory) teeth' must therefore be someone outside the control of the king or the diviner. The expression j]>L (loan for ^^) which can be interpreted as 'there may be / may not be a harvest', is also unmanageable since the bounty of a harvest is wholly dependent on uncontrollable factors in the natural environment. (For the semantic character of , see fn.l.) In the six hundred and twelve rubbings reproduced in Ping Pien, not a single tui-chen pair which divines about a controllable activity or event takes ch'i in both positive and negative sentences. - 153 -The phenomenon of the double appearance of ch'1 noted above can be explained in the following way, a way which is consistent with the traditional interpretation that ch'i is a modal conveying a sense of uncertainty. In the case of controllable activities, the Shang normally made a divination only after having decided either to perform or not to perform some activity like hunting or mobilizing troops. The outcome, at times in the human, at other times in the supernatural sphere, of this pre-determined course of action was the subject of their divinations. So the preliminary step to any divination about controllable activities was a decision either to take action or not to take action. It is precisely because the Shang, when performing a divination, are already inclined one way or the other towards the action, subject of the divination, that ch'i is employed in the sentences which express the unintended alternative. Since this alternative is not the one they are planning to take, to divine about its outcome becomes somewhat a matter of form to fit the tui-chen pattern and ch'i is employed to show their indisposition toward this choice. On the other hand, since they are planning, and are thus relatively certain of adopting the other choice, ch'i is not used in the sentence divining about the intended alternative. (Such an explanation, in fact, coincides with Serruys' insightful observation that ch'i is frequently associated with the less desirable alternative. Presumably and naturally, an unintended alternative is the less desirable alternative and vice versa.) - 154 -In the case of an uncontrollable activity or event, e.g., ^ j^g 'it will rain' vs ' ', since the Shang king and his diviners are quite incapable of knowing what the outcome will be, both possibil ities are equally uncertain. Thus the modal ch'i appears in both the positive and the negative divinatory statements. As to why ch'i is fre quently associated with the less desirable alternative, the answer may lie in the common psychology of human beings. People often refuse to anti cipate misfortunes and, once they have occurred, refuse to accept them. On first learning he has a malignant cancer, many a patient is apt to doubt the physician's diagnosis. In a similar way, the Shang king would not like to believe that his wife may die of sickness, nor to anticipate the possibility that no game may be caught on a hunting expedition. It is probably because of this mentality that the modal ch'i is frequently associated with the less desirable alternative to show the omentaker's unwillingness to accept an unwelcome possibility. Admittedly, this explanation is difficult to substantiate by means of factual data. Since the mood in a sentence is very subtle, without a larger context it is hardly possible to explain or to predict the employment of this presumed modal. However, since the theory of un-desirability seems to run into contradictions, and the modal usage of ch'i is well established in the classics, the explanation offered above seems to be a reasonable hypothesis. (.3) The interpretation of ch'i as a modal conveying the sense of uncertainty provides a convincing explanation for the following phenomenon. - 155 -The pattern ' (NP) + + ijif^ ' is commonly encountered in the O.B.I. For example, ting-hai / crack / test / Prince YU / perhaps / have sickness Ch'ien 5.44.2 Prince Yd, perhaps, has sickness. R Mi Lf* 9-?.'* test / Prince R (.?) / perhaps / have / sickness Prince , perhaps, has sickness. Ho 292 (Also Ch'ien fr) 4.32.2, Chia 3512, Ho & 208, I Z> 7817, Wen 841, Ch'ien 4.10.7, Ch'ien % To 1.335 and Ts'u tffi 1269. List inexhaustive.) However, with the rare exceptions of Ping 29 (2), Ping 31 (2) and Ts'un Fu ifyj' to which we will return, the structure ' (NP) + ^tl + ^ ' never appears in the subordinate clause of a simultaneous -successive sentence. Instead we find: ft* V% °* * test / have / sick / eye / favorable (So and so) has an ailing eye; he will recover, (lit. it will be favorable) . I_ ~Z-> 960 W I <±" * $ * GS * ft /' ' *• test / have / sick / tooth / not / wei / Father I .. JL hlrJ^ I Z 4626 (So and so) has an ailing tooth; it is not because of Father I's harm. - 156 -Or: CSo and so) has an ailing tooth; it is not Father I who causes the harm. have / sick / eye / not / prolong (So and so) has an ailing eye; (the sickness) will not be prolonged. Ho 210 have / sick / eye / perhaps. / prolong (So and so) has an ailing eye; (this sickness) will ^perhaps be prolonged. There are also cases of + body part, main clause', e.g., } A K fcm y* ft sick / trunk / not / wei / have / harm (So and so) is suffering from an ailing trunk; it is not (because of) there is harm (from the spirits). Ho ^ 441 test / sxck / mouth / exorcise / to Ancestral Mother Chia Ho 123 (So and so) is suffering from an ailing mouth; we should perform an exorcism to Ancestral Mother Chia. But the modal ch'i, with the exceptions mentioned above, never appears, i.e., the structures !*(NP) + jL + 4i- + , main clause' •*(NP) + JjL + >Ji + body part, main clause' are non-existent. 13 - 157 -Now, let us look at the three exceptions: king / prognosticate / say / Feng / perhaps go out / perhaps / wei / ting / not / go out perhaps / have / sickness / not / perhaps"* recover - Ping 31 (2) Feng will probably go out (to take the field) on a ting day. If Feng does not go out (to take the field on a ting-day), (it is because) Feng probably has an ailment and has probably not recovered. On the reverse side of this plastron, there are the following inscriptions: wu-yin / crack / Ch'Ueh / test / Hsien Feng perhaps / come Ping 30 (1) Hsien Feng will perhaps come. ft A*.^* T* test / Feng / not / perhaps / come Feng will perhaps not come. Ping 30 (4) These two inscriptions show that when the divinations were performed, Hsien Feng (abbreviated form: Feng) was probably in the field away from the Shang king and thus he was not under the direct control of the king (notice the *p- type negatives). - 158 -Ping 29 (2) is simply a duplicate of Ping 31 (2). These two inscrip tions are almost identical, the former has the crack numeral (hsu1 shu /j" ) 3 and the latter has 4; and the other inscriptions appearing on these two plastrons are also almost identical. The inscription on Ts'un Fu f|- P^j" 1. has some obliterated graphs, but the legible part is identical to Ping 31 (2) and Ping 29 (2). All three inscriptions are probably members of a ch'eng t'ao ( h\\ %^ set). Sickness is obviously something undesirable. And yet, ch'i only appears in the sentences or clauses which represent something the Shang king (or diviner) is uncertain of. In cases where the sickness has already taken hold, ch'i does not occur. Such a contrast also points to the fact that ch'i conveys a sense of uncertainty. For the above three reasons, the author argues that the traditional interpretation of ch'i as a modal conveying a sense of uncertainty is applicable to the O.B.I, language. In a subordinate clause ch'i may thus be translated by 'if. But according to this interpretation, ch'i cannot be a grammatical marker of conditional clauses. It is the semantic quality of its modality which allows the translation 'if in English. (See also Keightley 1978:66, fn.41.) 159 -2. An Evaluation of the Theory that Ch'1. Functions as a Subordination Marker Concerning Serruys' theory that ch'i is a marker of subordination, Takashima's article 'Subordinate Structure in Oracle Bone Inscriptions with Particular Reference to the Particle Ch'i ' deals with this problem in detail and demonstrates the difficulties arising from this theory. The author agrees with Takashima that 'the question of subordi nation must be determined essentially from context and sentence parallelism' (1977:38). The following are some brief remarks and discussion of residual questions: In general, the method of testing the validity of an interpretation of a presumed grammatical functor (or hstl-tzu %- in traditional terminology) is to apply one's interpretation to the several different contexts in which the presumed functor appears. In the case of ch'i, as Takashima has pointed out, it can appear in simple sentences (ibid.60). This phenomenon is not compatible with the theory that ch'i functions as a subordinate marker reflected in English translations by the words 'if', 'when' or 'rather than'. However, ch'i also functions as a modal. Thus, in simple sentences where it cannot function as a marker of subordination, ch'i might be interpreted as a word signifying a presumably less desirable alternative. Serruys has noted, "... it would be rather unusual if one and the same particle would be used for functions so strongly diverging, at least when viewed in the respective translations by which this particle has been expressed.' (1974:56) - 160 -In order to reconcile these two uses, Serruys gives the following explanation: 'since the "less desirable alternatives" are mentioned in the divinatory texts, they were, quite definite and factual, though undesirable, possibilities, and obviously they were so by the will of the gods or the ancestral spirits. If so, ch'i would simply stand for an adverbial term: "if (the gods) do wish so", i.e., "by divine fiat". This interpretation would permit Cl) the shift to sub ordination function in clauses of the type: "rather than doing a_, do _b," where the original ch1 i "less desirable" is still clearly preserved. (2) the function of double ch'i (as described above). (3) The function of ch'i in subordination before and after the main verb. (4) The development into the modal ch'i of later classical Chinese, which basically functions as an optative, exhortative and "probability" particle, a broadening of the original ch'i: "if (they) wish so".' (ibid. 58) The second point in the above analysis has already been treated on page 148-154. The first and third form the focus of the remainder of this section. Concerning the first point, there does not seem to be strong con textual evidence that would support the interpretation 'rather than ...' in the inscriptions provided and interpreted by Serruys. £ % " %L , ty fa "The king, rather than follow (someone) to take (prisoners?), he ought not follow to take (prisoners?). II The Fou rather than to bring for us together troops, they will not for us gather troops. it (Cited and translated by Serruys', 1974:36-37) - 161 -It is difficult to determine whether ' was, in fact, the preferred alternative to ' ^ hk 1. Neither can we speculate on the reason why the Fou would rather not "bring for us together troops". These interpretations are only based on the hypothesis that ch*i ^ conveys a sense of undesirability. On the other hand, the positive and negative counterparts of these two pairs are inscribed in perfectly symmetrical positions, i.e., they form typical tui-chen pairs. As a general rule, the two sentences of a tui-chen propose two complementary possibilities one of which will be identified by reading the oracle as the more likely to occur or to meet with supernatural approval. However, it would seem that only one possibility is seriously proposed in the above quoted tui-chen pairs. For example, in the pair % jt /w^ljx, 5- 'TH AK , the placing of the connective 'rather than' in front of the proposal 'to follow someone to take (prisoners)' suggests strongly that the Shang have already rejected this possibility. Of the possibil ities, one may be favored above the other; however, to use theconnective 'rather than' appears to exclude entirely one of the two possibilities. Moreover, although the two sentences are related and form a unit of divination, they were in fact divined separately, no doubt consecutively, so to treat them as two clauses of a single composite sentence seems rather anomalous. . - 162 -The third point in the analysis by Serruys cited on page 160 in fact involves two kinds of subordination as indicated by the placing v of ch'i before the main verb and after it. A. Ch'i before the main verb, a structure reflected by an 'if/when clause' in English translation. B. Ch'i after the main verb, a structure reflected by a 'that clause' in English translation. We will discuss these two points in succession, As to A, the presumed semantic-syntactic relationship between the sense of undesirability and the subordination marker 'if/when' does not seem to be very strong. The following inscriptions, among others, do not seem to provide a satisfactory support for the thesis. %. \t kf 5 * A: *ty When continuing to make exorcist rites for Ta-wu (then) we will have.a banquet. (Cited and translated by Serruys' 1974:49) yfl S « 1.1 0>tf)£ 2. #f next Yi-yu day, if we make offering (of wine?) to Tsu-yi, then we shall give a banquet. (Cited and translated by Serruys' 1974:52) In these, it is not clear why the sense of undesirability of ch'i should give rise to the connectives 'if/when'. Also there is no contextual! evidence to demonstrate that ' f f ' and ' JL ^ C^^f )' are - 163 -alternatives of the kind, 'if (the gods) do wish so'. Even if ch'i really conveys a sense of undesirability, it seem it can only stand for '(the gods), after all, do wish so.' The sense 'if/when' as Takashima argues, can only be apprehended through the context: £ * SSL* If the king yu-sacrifices captives, it will meet with (Ti's) approval. I * ft ft If the king yu-sacrifices captives, it will not meet with (Ti's) approval. (Cited and translated by Takashima 1977:45) Now, let us examine the following inscriptions: Ping 247 (1) (Cited on p. 145.) Lady Hao will give birth to a child, it will be good. The king prognosticated and said,~"(if) it is on a ting day (that she) gives birth to a child, it will be good; (if) it is on a keng day (that she) gives birth to a child, it will be protractively auspicious. (After) thirty one days, on a chia day, (Lady Hao) gave birth to a child, it was not good. (The child) was a girl. hsin-wei / crack / Ch'Ueh / test / Lady NU / bear / good / king prognosticate / say / perhaps / wei / keng / bear / good / three month / keng-hsU / bear / good Ping 257 (1) -164 -Lady NU will give birth to a child, it will be good. The king prognosticated and said, '(if) it is on a keng day, it will be good. At the third month, on a kerig-hsd day, she gave birth to a child, it was good. These two inscriptions appear to be evidence against the theory of the semantic-syntactic relationship between the 'undesirable' ch'i and the subordination marker ch'i. In both cases, clauses of the pattern + date + -v/L>' are subordinated to the main clauses ^7^'good' or J|! ^ 'protractively auspicious'. The theory of undesirability is inapplicable. Nor is the other interpretation of ch'i (i.e., as a subordination marker) appropriate here. If a possibility is deemed to be 'good' or 'protractively auspicious', it cannot, at the same time, be an eventuality the Shang would resign themselves to with 'if (the gods) do wish so.' As to B, the presumed semantic-syntactic relationship between the sense of undesirability and the post-verbal subordination marker (i.e., 'that' ....) would seem to be very "remote. In addition, to interpret.ch'i as a post-verbal subordination marker also encounters difficulties in terms of the general syntax of Chinese. Let us examine the following three groups of inscriptions: sue*) * . Ii tiS «• £ •* C The king prognosticating says: as to God, it is the - 165 -present second month that he will command there be hail (or sleet?); if it is on a Ping day that it will not (by his) command rain, then it means the Keng day will be auspicious. The king prognosticating says: it is auspicious that there be sleet (hail?). % . JIB © - t^t % ifv $ £ O h i The Hsi will bring white horses. The Hsi will not be bringing white horses, amounting to five. The king prognosticating says: it will be auspicious that they come. (The order of the translations of this group of inscriptions is altered. The translation of the negative counterpart is inserted between the translations of' the positive counterpart and the prognostication.) ^ ffl • t Pi. £ .* ii ! The king prognosticating says: it is auspicious that when Yin arrives, it will be a Hsin day. (Cited and translated by Serruys', 1974:34) Most scholars interpret the word chi ^ in the structure ' JL I3) ^ : ' as a one word sentence, i.e., the king prognosticated and said, "(it is) auspicious." (As an example, Chang Ping-ch'Uan adds punctuation as follows: 1> fj*l \3 : %• I jt- ^,(1957: Vol. 1.1, p.99); £l9 S : * , -£fc.<1959: Vol. 1,2, p.232); £ ]§ VQ : £ % (Ibid. p. 144). - 166 -Although Chang does not use punctuation consistently, there is no doubt he considers a syntactic break or even a sentence boundary to come between the word chi ^ and the words following.) However, this kind of parsing cannot be adopted here for doing so would yield sentences of strange meaning: *The king prognosticated and said, "it is auspicious". It will (undesirably) hail (sleet?). *The king prognosticated and said, "it is auspicious. (So and so) will (undesirably) bring in *The king prognosticated and said, "It is auspicious. Yin will arrive, (undesirably) it is on a hsin day. (All translations the author's) The function of ch'i jt, as an indication of the undesirability of an activity/event has already been discussed above (see p. 142-158 ). If it does not seem entirely applicable in a given case, it might be argued that ch'i loses its presumed affective meaning because of the existence of chi rz, and becomes simply a subordination marker, translateable as 'that ...'. However, this interpretation also encounter difficulties. In the second example: - 167 -The expression %^ appears both in the negative counterpart and the prognostication. Serruys' analyses of these two ' J% ' are different. In the prognostication, ' J% ^ ' is analysed as 'subordination marker + clause' while in the negative counterpart ch'i ^ seems to be interpreted as a word signifying undesirability. The use of the same element in two different ways within the same inscription is not a priori impossible, but it may be suggested that the focus of this group is whether or not the Hsi will arrive with their tribute (?). However, to interpret the prognostication as 'it will be auspicious that they come' (the author would suggest understanding it as a causative verb 'to cause to come — to bring in') conveys the impression that the Shang are divining about whether the 'bringing in' is auspicious or not to which the response is: 'the bringing in is auspicious'. In order to maintain this interpretation, we would have to assume a shift of divination focus. Another way of interpreting ' as an embeded sentence (i.e., a 'that clause') is to take ^ as a putative verb, i.e., 'to consider the bringing in to be auspicious'. However, such an interpretation shows even more clearly the shift of divination focus indicated above. The last inscription ' "j? J^t /j^ ^ ' poses another problem. Here, if ch'i is being interpreted as a subordination marker placed after the clause fpT , the word for word translation is: '*It is auspicious Yin arrives that will be a Hsin day'. In Serruys' translation -'it is auspicious that when Yin arrives, it will be a Hsin day', ch'i is being analysed as governing the entire structure '* ? ^ ^ '• Such - 168 -an interpretation seems to violate the general syntax of classical Chinese. This sentence exemplifies the problems arising from the interpretation of ch'i as a subordination marker. A final example further illustrates the inappropriateness of both the subordination marker theory and of the undesirability theory: king / prognosticate / say / auspicious / ti-god W 1 £ 1 t £ A x perhaps / give / I / assistance Ping 410 (1) i. Using Ch'i as a subordination marker yields — *It is auspicious ti-god that will give me assistance. ii. Using Ch'i as a word signifying the less desirable alter native yields — *It is auspicious. Ti-god will give me assistance (a less desirable alternative). Returning to the first inscription (see p. 164), it is easy to see that the interpretation of ch'i in the structure ' <^ «• ' as a subordination marker is rather unnatural. This interpretation requires us to take this sentence as a 'determinative sentence', i.e., a is b, ' ' would be the 'determinated' while ' «, ' is the determinator, i.e., '*keng day is auspiciousness' or '*it is a keng day that it is auspicious'. Either interpretation involves unnecessary' complications. Neither can we interpret ch'i as a word signifying the sense of undesir ability, since the translation '(if) it is a keng day, it will be (undesirably) auspicious' is rather strange. - 169 -This analysis of the logical meanings and syntactic structure of the above inscriptions leads us back to the traditional interpretations ' £ k • J- t (If) it is a keng day, it will perhaps be auspicious. It is auspicious, (Hsi) will perhaps bring in (white horses). It is auspicious. Yin will arrive, it will perhaps be on a hsin day. Serruys also cited another type of post-main-verb subordination, e.g.« J^; 1^9 which is translated as 'not have (chance) that it might rain'. Serruys states, 'As Takashima observed the pattern wang ch'i t is used throughout with few instances of J^-* "t-as would be expected when compared with all other cases where ch'i precedes the verb as in _JL^ j: "there will be baneful influence," etc. The hypothesis presented to explain this exception is simply that wang ^C* is treated as a main verb "not have" followed by what is really an object clause, ..." (1974:57) Admittedly, according to the best of the author's knowledge, a satisfactory explanation for this peculiar structure is lacking. Unfortunately, the above hypothesis does not seem to shed much light on the matter. This hypothesis fails to explain why the structure never occurs. (The tremendous number of rubbings makes a thorough check, very impractical, if not impossible. However, in the Ping Pien where six hundred and twelve fairly complete plastron rubbings are collected, and in the thousands of other inscriptions the author has - 170 -encountered, such a structure does not occur.) Instead one does find the structure ' ^ iTO ' . Why is ytt ejfo made into a presumed embedded sentence (I.e., subordinative clause) after the verb wang -ii, but not after the verb yu 7^ (. ^ , )? As a matter of fact, the structure is also very rare while the structure ' -^7 lv|*)' is very common. A hypothesis proposed to explain this phenomenon is that ' 4i VvT9 ' is almost always, though not exclusively, used to represent the intended result of rain-soliciting-ritual-sacrifices. The negative counterpart of this structure is 1 •£ jL. ' which is used to represent an undesirable possibility. The employment of ch'i which conveys a sense of uncertainty reveals the unwillingness of the Shang to believe the undersirable possibility (see p. 154). While still failing to explain why ch' i is placed after wang XL instead of before it, the above hypothesis seems to be more defensible. - 171 -IV. A HYPOTHESIS ON THE RATIONALE UNDERLYING THE CHOICE BETWEEN THE PATTERNS 'V % I j^y .. . ' AND. ' JQ + V . . 'V .... , n In divining about the appropriateness of performing a certain action, the following two types of composite sentences are commonly used: 1. 'V % I tyo ... ping-shen / crack / Chung / test / call upon / Chien Ch'ien / * (?) / cut (grass) / net • in Ping 126 (6) If we call upon Chien, Ch'ien and (?) to cut grass, they will get (lit. net) it. Or: We called upon Chien, Ch'ien and J$- (?) to cut grass, they will get (lit. net) it. n*$> r >• % ping-shen / crack / Chung / test / call upon / Chien £ * * 8| Miff Ch'ien / & (?) / cut (grass) / not / perhaps / net Ping 126 (5) k If we call upon Chien, Ch'ien and 4£ (?) to cut grass, they will perhaps not get (lit. net) it. x Or: We called upon Chien, Ch'ien and (?) to cut grass, they will perhaps not get (lit. net) it. - 172 -2. ' ,l0 + V . .., % I ..." 1 v \hi At %t hsln-yu / crack / Ch'Ueh / test / this / season / king allow / Wang Ch'eng / attack / Hsia Wei /receive / abundant follow % assistance Ping 12 (1) The king should follow Wang Ch'eng to attack Hsia Wei this season, (for) he will receive abundant assistance. king ¥?$9 ri- Mi At |i jt hsin-yu / crack / Ch'Ueh / test / this / season / kinj Id not / follow / Wang Ch'eng / attack / Hsia Wei should #i yjt ^.f ^ * not / perhaps / receive / abundant / assistance Ping 12 (2) ^The king should not follow Wang Ch'eng to attack Hsia Wei, (for) he will perhaps not receive abundant assistance. In fact, both of these structures seem to serve the same divinatory purpose. When the 'responses' to the positive counterparts (the positive charging statements) are positive, i.e., *right, what is being charged is correct', the illocutionary force of Ping 126 (5) and Ping 12 (1) would be '*go ahead to do such and such'. Or, if the 'response' to the negative counterparts are favorable, i.e., '*right, what is being charged is correct'; the illocutionary force of Ping 126 (.6) and Ping 12 (2) would be '*don't do such and such'. Do these two patterns represent merely a difference of styles, or are there other factors determining the choice between them? The author has two hypotheses on this question. - 173 -1. The pattern ' ^ffl + V ..., / -tp ...' is used when the proposed activity has not been carried out and is under the direct control of the Shang king. The pattern 'V ..., ^ / ^ ...' is used when the proposed activity has been carried out or it is not under the direct control of the Shang king. The first part of this hypothesis seems very natural and acceptable. Sentences similar to Ping 12 (1) (2) incorporate the particle wu ^7 in the negative counterpart, and a presumed hortatory 'should' in the positive counterpart. Either prohibition or obligation, in most cases, presupposes the controllability/manageability of an action. A sentence such as 'go / don't go to the mountains' is perfectly acceptable providing the addresser has some sort of authority over the addressee. But a sentence such as 'be young', even if it is acceptable, can only be used in a very restricted sense and situation. Though wu shows a somewhat complicated character, its predominant function is undoubtedly a prohibitive negative whether in the O.B.I, or in the classics. While the pattern ' ^ + V ..., ^?v/^...' cannot be used in situations where the proposed action has already been carried out nor in situations where the proposed action is not under the direct control of the Shang king, we cannot prove that the pattern 'V ..., ?x / ^? ...' is used exclusively in situations where the pattern cannot be used. The next example is open to two interpretations: this / day / hunt / not / perhaps / net . / captive I ZJ 143 - 174 -Cl) (So and so) is hunting; he will perhaps not capture captives. That is, somebody has already set out on a hunting trip, and the Shang king (or the diviner) makes this divination to learn the result. This is a situation where the ' ^ + V / J^j type sentences cannot be used. (2) (If) we hunt today, we will not perhaps capture captives. That is, the Shang king (or the diviner) divines about the outcome of a proposed activity to determine whether he should go ahead with it. This is a situation where the ' 4p+v .... .' type sentences can be used. As illustrated on p. 118 and 142, in many cases it.is difficult to distinguish conditional sentences from simultaneous-successive sentences. Since many divinatory sentences like I ZJ 143 above do not include time expressions of any kind, it is often impossible to determine the temporality of the activity being divined about, i.e., whether it had already occurred or was in process and thus beyond the control of the Shang king, or whether it was a proposed action to be undertaken at some future time. For this reason, the hypothesis concerning the difference between the "^V type of sentence and the 'V ^ / ^> type of sentence is difficult to verify. There is a further difficulty: some 'V / jrjj~> type sentences appear to be used in situations where the activity divined about is controllable and not yet carried out. - 175 -< /it W| Ift m ^ ff 4%, king / perhaps / go / chase / game(?) I at I Kuei capture Ping 216 (5) (Ef)the king perhaps goes to chase game(?) at Kuei, he will capture (them). Or: (if)the king will perhaps go to chase game(?) and 14 kuei-deer(?), he will capture (them). t£ a* I* $ ff <f 4 king / perhaps / go / chase / game(?) / at / Kuei ^ W£ not / perhaps / capture Ping 216 (6) (if)the king perhaps goes to chase game(?) at Kuei, he will not perhaps capture (them). Or: (If) the king perhaps goes to chase game(?) and ^ kuei-deer, he will perhaps not capture them. The word ch'i appearing in front of the verbs 'frt -^L'to g° chasing'; shows that this activity has not been carried out at the time of the divination while the subject wang jt- 'king' shows that the activity is under the direct control of the king. f N* iii U£ \A %i V test / king / perhaps / chase / rhinoceros / capture not / retreat / rhinoceros / capture / pig / two (If) the king perhaps chases rhinoceroses, he will capture (them). The king did not drive a rhinoceros (into an enclosure) (lit. to make it retreat into an enclosure); he captured two pigs. Ping 120 (15) H y*, H not / perhaps / capture / rhinoceros (If the king perhaps chases rhinoceroses), he will perhaps not capture (them). Ping 120 (16) • H^'lV ^ h US %,± fj% ping-hsu* / crack / king I •we I perhaps / chase / game capture / indeed / capture / ten (If) we perhaps chase game, we will capture them. Ping 323 (1) ping-hsu* / crack / king / not / perhaps / capture / game < - 9 first month Ping 323 (2) (If we perhaps chase game), we will perhaps not capture (them). ^ (Divined) at the first month. In these two pairs of inscriptions, the modal ch'i _j£ is also used.^ And the agent or the diviner of the activity chasing is the king himself. Obviously chasing is something under the direct control of the king. Thus, in the same kind of situation, both types of sentence can be used. Is there any other factor that might determine the choice between them? Here is a second hypothesis: 2. The choice is determined by the gravity of the consequence of a proposed activity. Although a clear cut distinction between these two patterns is lacking, the author has discovered that certain verbs only appear in one pattern but not the other. For example, the verb fa A"^ , when used in - 177 -the sense 'to attack' (identified by the name of a tribe following it as direct object), appears only in the pattern ' ^ + ^ ^5"' / ^ ...', #«#«r ^ Hi fc-*> ^ kuei-mao / crack./ Ch'Ueh / test / call upon / Ch'Ueh (?) / attack / HsUan / harm Ping 249 (4) We should call upon Ch'Ueh to 4$f (?) attack the HsUan 16 (Statelet), (for) we will harm them. should not / call upon 7 Ch'Ueh / ^J(?) / attack &t #f yr* HsUan / not / perhaps / harm Ping 249 (5) We should not call upon Ch'Ueh tof^j(?) attack the HsUan (Statelet), (for) we will perhaps not harm them. der nTft* . ^ fir M* 1 ting-wei / crack / Pin / test / should not / ord Pi / attack / Hu(?) Statelet / not / perhaps / receiv ive abundant / assistance We should not order Pi to attack the Hu(?) Statelet, (for) we will perhaps not receive abundant assistance. I {'^ 17 crack / Cheng / test / this / season / king / follow 1 assi ,1k (fflj^TlA t\ * i eng / attack / Hsia Wei / receive / abundant Wang Ch" sistance HsU,tf|[ 3.11.5 - 178 -^ This season the king should follow Wang Ch'eng to attack Hsia Wei, (for) he will receive abundant assistance. 7*ft i > % nk hsin-ssu / crack / Cheng / test / this / season ± * *> tU< f * f * if * king / should not / follow / Wang Ch'eng / attack °* «i f% ^ * Hsia Wei / not / perhaps / receive / abundant / assistance HsU ^-f? 3.11.5 • This season the king should not follow Wang Ch'eng to attack Hsia Wei, (for) he will perhaps not receive ^•abundant assistance. test / king / hui / Chih Chia / follow / attack / Pa Statelet A* ti-god / give / we / assistance It should be Chih Chia that the king follows to attack the .J Pa Statelet, (for) ti-god will give us assistance. I_ Z«3787 king / should not / wei / Chih Chia / follow / attack Pa Statelet / ti-god / not / we / perhaps / give 1 assistance i_ 2/3787 It should not be Chih Chia that the king follows to attack the Pa Statelet, (for) ti-god will perhaps not give us assistance. - 179 -With the exception of J. ~Z> 3787 of which the diviner is not recorded, all the diviners, i.e., Ch'tieh^^ , Pin and Cheng 3^. are.different. • in these four groups of inscriptions. And yet the pattern ' + 4^ ..., . is maintained. Thus, the choice of this pattern can hardly be accounted for by assuming idiosyncratic use on the patterns on the part of the diviners. Although the negative counterpart is sometimes, in other inscriptions, abbreviated to .', we have not found any instance of the pattern A similar phenomenon is observed in the case of cheng CE to send a punitive expedition) . The pattern 1 %~) + Ct ... , ^ / -j^? ... 1 is also predominant: %jkl* 1 * it fs^ keng-shen / crack / Ch'Ueh / test / king / should not • ? t punitively attack / Hu(?) Statelet / lower / upper #* it?A-a* 4* ^* not / approve / not / we / perhaps / give / assistance Ch'ien- 5.22.2 The king should not punitively attack the Hu(?) Statelet, (for) the lower and upper (spirits) will not approve and (they) will perhaps not give us assistance. . . . yju / crack / test / king / punitively attack Hu(?) Statelet / lower ./. upper / approve / give / we 5^ assistance T'ieh $ffc .244.2 - 180 The king should punitively attack the Hu(?) Statelet, (for) the lower and upper (spirits) will approve and give us assistance. test / should not ,5, ±<ar?\ ^ I punitively attack Hu(?) Statelet / lower .-:/•'. upper / not / approve *\ -fc * X not / we / perhaps / give / assistance T'ieh-.jgjr, 244,2; (The king) should not punitively attack the Hu(?) Statelet, (for) the lower and upper (spirits) will not approve, and will 'perhaps not give us assistance. (See also I ^^116, Hayashi Z.9.6, Hou Shang b- 16.8, I ^18 and Ch'eng 354.) There is only one inscription which takes the pattern ' fjt-.. j $s . kuei-wei / crack I M (?) / test / king punitively attack / not / approve If the king punitively attacks, there will be disapproval. K\i 1177 On the other hand, in the cases of chu £^ 'to chase (animals)' and shbu 'to hunt', both patterns occur: (.1) Chu a. ' 4?+ /jW - 181 -test / next / tisin-ssu / king / should not / go chase / rhinoceros / not / perhaps / capture T'ung |J*/XI 3. On the next hsin-ssu day, the king should not go chasing rhinoceroses, (for) he will perhaps not capture (them). ft k ' £ * t> I * ^ * ^4 test / should not / call upon / chase/ not / perhaps S\ f capture^ I £J 7600 We should not call upon (so and so) to chase (animals), (for) we will not capture them. g / kin  / perhaps / chase /pp-game/ at / Se * 17 capture(?)i' I ^ 7490 The king should perhaps chase p*^ -game at Se, (for) he will capture (?) them. should not / chase / p^f -game / not / perhaps capture _i 2^ 7490 The king should not chase -game,"(for) he will perhaps not capture (?) them. should not / call upon / chase /... / perhaps / ... We should not call upon (so and so) to chase (animals), (for) we will (not) perhaps ... Chin - 182 -(In T'ung ^ IX 3, I 7600 and Chin ^ 404, the verbs being negated by wu ^) axe either wang or hu ^ (=t?^ ). However, in the underlying rationale, the focus of negation is the hunting activity chu )%\ . Thus, these inscriptions are listed under the heading of chu l^-.) „ / 1. / / -1 / / 1 Ts'ui 936 chi-mao / crack / king / chase / rhinoceros / not . If the king chases rhinoceroses, he will not ... . (See also the three examples cited on p.175-176 .) *T'to hunt' test / king / hunt / :net -• Ping 423 (4) The king should hunt, (for) he will capture (animals). ... wei I ... I jen-shen / king / should not / hunt not / perhaps / net / jen-shen1 I ... I hunt / net' Ping 423 (2) On the jen-shen day, the king should not hunt, (for) he will not capture (animals). On the jen-shen day, (the king) went hunting and captured (animals). - 183 -suu / crack I ... I king / should not / hunt / ... not / please Liu Chung -f, 106 The king should not hunt, (for) ... will not be pleased. keng-hsu1 / crack / this day / hunt / not / perhaps . net / captive I ^ 143 (If) we hunt today, we will perhaps not capture captives. hsin-mao / crack / Cheng / test / we / hunt Hsia I / not / please HsU \\ 1.46.3 (If we) hunt, Hsia I will not be pleased. It appears that an important action which may have serious consequences in the Shang people's mind is more likely to motivate the employment of the pattern ' 4%) • • • , "7* /••^•j* ...'. 'To attack a statelet ^[\^ , ^i- ' is, needless to say, an activity the success or failure of which has great impact on the Shang. Divinations concerning this activity almost invariably take the pattern ' ^) /?*/ -|? ...'. On the other hand, the success or failure of a hunting activity can hardly be compared in :. importance to that of a military expedition. We find that both structures .' and 'V > I'...' are used in the cases of chu and shou ffi . - 184 -The verb t'ien I+? 'to hunt', which does not often occur in period I O.B.I, seems to be an exception. It frequently, appears in the pattern ' ff^ (='//] )©..., J^- 'flf " (55 occurrences) 'We should not hunt... for it will be cause for regret', though seldom in the pattern ' |5 ..., 0^ ...' (10 occurrences) if we hunt ..., we will not ...'. Although the pattern ' yj (= W) (±f % ' is different from that of ' ^ ... , $ £ >^ the expressions ' $ ?f « and ' t £ ' both convey a negative sense. And both patterns express a 'cause and effect' relationship. In this respect, the verb t'ien |2 'to hunt' is similar to the verbs fa , cheng and distinct from the verbs shou ft , chu . However, it should be noted that it is not the possible failure to capture game that constitutes the major reason for not going hunting ( (if ), in fact, there is only one inscription which reads ' [±I .. ., 7j* 3^. '' •1 • Instead, it is the possibility of 'cause for regret' f-^ . Cf. 7L % fa '-flf (I Ching HY 1/1/ jl, fa ) 'Arrogant dragon will have cause to repent.' (Baynes 1950:8) In the O.B.I, and in later texts, there are passages which specify by concrete examples just what 'cause for regret* meant to the Shang. U #W Ms b% It kuei-ssu / crack / Ch'Ueh / test / ten-day-week / have no misfortune / king / prognosticate / say / nai / this / also have / harm / similar to / say / chia-wu / king / go 185 -chase / rhinoceros / small / official / to take charge chariot / horse / precipitous / drive / king / chariot f ±^ /If * Prince J (?) / also / fall T'ung (|/_ 735 There will not be misfortunes in this ten day week. The king prognosticated and said, 'here will be harm too.' (It turned out) just as the king had prognosticated. On the chia-wu day, the king went to chase rhinoceroses. A minor official took charge of the chariot and horse, he drove the king's chariot precipitously. Prince ^ 18 also fell down. Shih Chi: $t tf) vf i 1^ , %L % . $L (Yin Pen Chi jf£ ^ , chuan 3, p. 104) 'During a hunting trip between the Yellow River and River Wei, Wu I was killed by a fierce thunder strike'. In the Han Dynasty, Ssu-ma Shang-ju admonished the emperor not to go hunting because of the dangers involved. He states: M * ^vi fo.K* *J • • • • ja. * & fit tir . * ^ ^ &«t > t% - 186 -(Ch'Uan Shang Ku San Tai Ch'in Han San Kuo Liii Ch'ao Wen , P.246) Now your majesty is fond of surpassing the obstructions and dangers, and shooting fierce animals. (Supposing your majesty) suddenly encounters an animal which has extraordinary strength (at the time when) your majesty least expects it. (It) attacks your majesty (lit. it offends the clear dust of your majesty's chariot), but there is no time to turn the chariot around and no time for your majesty's people to use their skills (to protect your majesty) ... (Such a situation is equivalent) to having the Hu and YUeh (barbarians) rising from the bottom of your majesty's chariot, while the Ch'iang and I (barbarians) are on your rear bumper. Isn't it perilous! ... Moreover, even if one drives after the road has been cleared (from obstructions) and drives fast down the middle of the road, occasionally there are chariot accidents; not to mention if one gets into bushes and drives fast on uneven places ... isn't it difficult (to avoid) the misfortunes! This kind of mishap may be what the Shang had in mind when they consequences of such misadventures are much more serious than the mere failure to capture animals. This may be the reason why the pattern interesting to note that in period I inscriptions, the diviners rarely divine about the possibility of these mishaps (e.g., misfortune) and the major concern of hunting divinations lies in whether or not animals performed the divination Needless to say, the It is will be captured. Luxurious Ease) Chapter of the Shang Shu, the Duke for his diligence and reprimands the' Shang kings who followed for - 187 -their luxurious life-style in which extravagant hunting must have played a part. However, there are many hunting inscriptions from Period I, and thus presumably hunting was an important court activity, yet Wu Ting is not singled out for extravagance by the Duke of Chou. A possible explan ation is that hunting in Period I (the reign of Wu Ting) was a practical 19 necessity, i.e., a vital means of obtaining food and sacrificial victims. (See Yang K'uan's Ku Shih Hsin T'an ^_J^f i$>» P-259-260). This may be the reason why the capturing of game was the predominant concern in Period I hunting divination. Hence, if the criticism aimed by the Duke of Chou has any historical validity, we might assume that in the later periods hunting was carried out mainly for pleasure. Thus, concern about the success of the hunting gradually gave way for the king's safety in what had become little more than a dangerous sport rather than an essential - 188 -CHAPTER FOUR THE SUBORDINATIVE COMPOSITE SENTENCES II (RITUAL - SACRIFICIAL VERBS) The term 'ritual-sacrificial verbs' (hereafter R-S verbs), refers to verbs which take supernatural beings, e.g. the spirits of deceased ancestors, mountain (gods), etc., as indirect objects identified by the co-verb vjl 'j" . (There is also a case where an R-S verb, i.e., pin /& 'to treat as guest', takes a supernatural being as direct object.) The co-verb y_U does not always occur between an R-S verb and its indirect object, but, in any cases where the pattern 'V + -f + supernatural being' occurs, the verb in question can be classified as an R-S verb. Various investigations have been made on R-S activities (chi-ssu ^ X\L>) since the earliest stage of O.B.I, studies. Among them, Ch'en Meng-chia's 'Ku Wen Tzu Chung Chih Shang Chou Chi Ssu' t fc. % <f % ffi (1936) and Shima Kunio's Inkyo Bokuji Kenkyu f'^ ^ [x fff £ff'^(1958) are the most comprehensive. The major concern of these scholars is with the purposes/functions, the character, the beneficiaries (i.e., indirect objects) of these R-S activities, and frequently, their identification with similar ritual activities recorded in the classics. In this chapter, it is not the intention of the author to re-discuss these R-S activities or to examine the theories proposed by the previous scholars. Instead, the focal point lies on the syntactic relationship of the clauses in which R-S verbs appear. Only in cases where the author finds the previous interpretations of a - 189 -given R-S verb unsatisfactory will he discuss the possible meaning of that R-S word. Otherwise, he simply follows the generally accepted opinions. The syntactic relationship between two R-S clauses is, in fact, a reflection of the relationship between two (or more) R-S activities in the order of those activities, i.e., a major R-S activity in relation to a subsidiary:R-S activity. In this chapter, an attempt is made to illustrate such relationships by studying the different patterns exemplified by different R-S verbs. The more inscriptions we investigate, the more definite a generalization we can make. Therefore, we select as objects of study R-S verbs which have a high frequency of occurrence. The observation of these verbs is based on the data, regardless of period, collected in the Sorui. Although failing to discriminate data of different periods may blur the temporal changes of these verbs, it would be impossible, without a large number of inscriptions to arrive at any solid conclusion. Therefore, a wide-ranging treatment seems to be necessary. One finds traditional terms such as chi-ming and yung- sheng chih-fa used to refer to different types of R-S verbs. (The Japanese scholars, such as Shima (1958:258) and Ikeda (1964:1), use terms saimei ^Jfc , saigi tjf^ , saishi yogo jj£ 3& $5 via..) However, these terms have not been clearly defined. In this chapter, the author attempts to give a linguistic description of these verbs. They are described simply as 'type A' and 'type B'. These two types of verbs can be clearly distin guished syntactically in certain cases as will be shown in the following passages. In other cases, however, we have to rely on the method of analogy and comparison in order to classify a verb. For the present, the - 190 -following three patterns are selected as a paradigm. Of these, (2) and (3) are used to illustrate the distinction between the type A and type B verbs. (.1) 'V + (A ) + OB' (2) 'V + C "f ) + OB + OV' (3) 'V + OG/OP + C "? ) + OB' (See below for the meaning of the abbreviations used in the paradigm.) Of course an R-S verb may appear in many other patters, but it is impractical and unnecessary to discuss them all. The above three patterns can serve our purpose. Abbreviations: OB object-beneficiary: the recipient of an R-S activity, e.g. a deceased ancestor OV object-victim: the animals, things being sacrificed. There are two sub-categories of OV depending on whether it is used with verb type A or verb type B, 01 object-instrument: an OV in relation to a type A verb, e.g. an ox OP object-patient: an OV in relation to a type B verb, e.g. an ox 0G object-goal: the goal or cause of an R-S activity, e.g. (for) a good harvest (see footnote no.3) - 191 -I. A GENERAL DISCUSSION OF THE TYPE A AND TYPE B VERBS The contrast between Type A and Type B verbs: (.1) V + (. -j- ) + OB Type A test / king / announce / to / Ancestor I The king should make an announcement to Ancestor I I Z> 8134 test / should not / announce / to / T ang We should not make an announcement to T'ang. Ch1 len #'J 1.47 I should not / exorcise I to I Ancestral Mother Chi We should not perform an exorcism to Ancestral Mother Chi. I ZJ 5406 test / exorcise I to I Ancestral Mother Chi We should perform an exorcism to Ancestral Mother Chi. I Z> 5406 test / invoke / to / River(god) HsU y^. 1.35.4 We should invoke to the River(god). 3ii £fc i £ f-r t *ct ^ should not / invoke I to I Ancestral Mother Keng. We should not invoke to Ancestral Mother Keng. I 7448 - 192 -e B jen-shen / crack / Pin / test / burn-sacrifice to / River(sod) Chia vf 2998 We should burn-sacrifice to the River(god). test 7 should not / burn-sacrifice I to I River(god) We should not perform a burn-sacrifice to the River(god). Chia Vp 2998 5i J | f& ^ ^0 t f hsin-ssu / crack / yu-sacrifice / to Ancestral Mother Ting I_ Z 4064 We should perform a yu-sacrifice to Ancestral Mother Ting.1 kuei-mao / crack / Ch'Ueh / next / chia-ch'en yu-sacrifice / T'ai Chia Ping 117 (22) On the next chia-ch'en day, we should perform a yu-sacrifice to T'ai Chia. test / come / i-hai / yu-offer / to / Ancestor I On the coming i-hai day, we should perform an yu-of f ering to Ancestor I. Ping 313 (1) 6ii'j=U J^t* -: v ^ ^ ,:h test / come / i-hai / should not / yu-offer / to Ancestor I Ping 313 (2) On the coming i-hai day, we should not perform an yu-offering to Father I. - 193 -(2) V + C f ) + OB + OV Type A * 11 ru)!* U ft *U °x ping-shen / crack / announce / to / Father Ting ox / one Nan Ming ^ 531 We should make an announcement to Father Ting (with) one ox. chi-hai / crack / announce I to I Father Ting hree / ox thre x ' Nan Ming 616 We should make an announcement to Father Ting (with) three ox. ting-ch'ou / crack / Cheng / test / exorcise / to Ancestor Hsin / ten / penned-sheep We should perform an exorcism to Ancestor Hsin (with) ten penned-sheep. Hou Shang ^ Jz 27.1 211 A ,U, 2, ... exorcise / to / Nang King / three captive We should perform an exorcism to Nan King (with) three captives. K'u j^L 1641 Hfi °T h U fi D'> ^ test / next / ting-hai / invoke / to / Ting / two / ox On the next ting-hai day, we should, invoke to Ting (with) two oxen. Pu \ 244 -194 -3ii j^fV 1 K ?f M| I & fj i-hai / crack / Cheng / test / invoke / to is * u y* * Hsien / ten / ox Ch'ien fi'j 1.44.2 We should invoke to Hsien (with) ten oxen. Type B test / burn-sacrifice I to I Ting / five / ox We should burn-sacrifice to Ting five oxen. Ch'ien 0(J 1.46.5 4ii f«pfl* At ^W^C chia-ch'en / / king / next / i-ssu burn-sacrifice / to / Hsien / one / sheep On the next i-ssu day, we should burn-sacrifice to Hsien one sheep. I_ Z. 849 test / yu-sacrifice / to / River(god) / ten / ox We should yu-sacrifice to the River(god) ten oxen. I Z. 7645 5ii f $ f * $ # Ts I* / + hsin-wei / test / yu-sacrifice / Shih Jen / ten penned-ox Nan Ming $£)Pfl533 We should yu-sacrifice to Shih Jen ten penned-oxen. next / hsiri-hai / yu-of fer / to / Wang Hai forty / ox Ping 117 (.31) On next hsin-hai day, we should yu-offer to Wang Hai forty oxen. - 195 -6ii y * f ? tj. f % £ i yu-offer / to / Wang Hai / female slave We should yu-offer to Wang Hai a female slave. Ping 117 (30) (3) V + OG + ( f ) + OB Type A 4 u ft ft £i&*-tf*f*[±U » test / announce / T'u Statelet / to / (Shang)Chia We should announce (the activity of) the T'u Statelet to (Shang)Chia. Ts'ui ^ 1107 test / announce / Hu(?) Statelet / to / T'ang We should announce (the activity of) the Hu(?) Statelet to T'ang. HsU 1.7.2 exorcise / king / misfortune / to / Ancestral Mother Kuei I_ Z. 7781 We should exorcise the king's misfortune, (the 2 exorcism being directed) to Ancestral Mother Kuei. 2i.i M*r ±* ff should not / exorcise / king / misfortune / to Ancestral Mother Kuei I_ Z> 7781 We should not exorcise the king's misfortune (the exorcism being directed) to Ancestral Mother Kuei. - 196 ->fc '31 %> & ^ % invoke / harvest / to / River(god) We should invoke for a (good) harvest to the River (god). Hayashi ^2.19.8 3ii ff^H i K # * ™<w $ 4f keng-wu / crack /invoke / rain / to / Mountain(god) We should invoke for rain to the mountain(god). I 40 (3) V + OP + ( ^ ) + OB Type B 14i u i- % 4 - ^ chi-mao / crack / burn-sacrifice / pig / four cloud K|_u /£ 972 We should burn-sacrifice a pig to (the spirits of) the four clouds. (For a study of cloud-worship, see YU Hsing-wu 1979:8.) 4n Ufi # I Xfc ijfc. ""1 i-ssu / / burn-sacrifice / five / pig / one sheep / to / ^ £ (?) Jin A. 2299 We should burn-sacrifice five pigs and one sheep to the spirit of ^ (?). i, 51 A* 08 .,• test / this / day / yu-sacrifice / small / penned-sheep / to / Father I HsU fyT 2.21.4 Today we should yu-sacrifice a small penned-sheep to Father I. - 197 -ting-hai / crack / at/next / wu-tzu / yu-sacrifice Hiffc & ah "fatJ. three / pig / Ancestor I / keng-yin / use four / month HsU %% 1.13.3 On the next wu-tzu day, we should yu-sacrifice three pigs to Ancestor I. (This sacrifice) was performed on a keng-yin day (instead). On the fourth month. W 6i Wy* / k T> '; *ft ^ # yu-offer / captive / to / Ancestral Mother Keng We should yu-offer a captive to Ancestral Mother Keng. Ping 47 (21) 6ii ft ^ * * f * f" «tf should not / yu-offer / captive / to / Ancestral Mother Keng Ping 47 (22) We should not yu-offer a captive to Ancestral Mother Keng. At first sight, it appears that all the verbs listed under A and B share identical linguistic behaviour. All of them can take an OB, i.e., pattern (1)'V + ( ) + OB'. Other than the OB, these verbs can also take another object which is either placed in front of or after the OB, i.e., pattern (3)'V + OG/OP + ( y ) + OB' and pattern(2)'V + ( -f ) + OB + OV'. Nevertheless, it is easily observable that in pattern (3), the objects of the 3 type A verbs are OGs while the objects of the type B verbs are OPs. - 198 -One may hold the opinion that both an OG (e.g. T'u Fang S~ Jj ) and an OP (e.g. ox ij' ) are direct objects; and the difference between them is a matter of meaning devoid of syntactic significance. It is a common knowledge that, in classical Chinese, verbs such as y_u ^ 'to give' and kao 'to tell' take two objects (direct object and indirect object (or OB)). The direct object is the thing/matter being given/told, while the indirect object is the recipient/listener of the thing/matter. The direct 4 object may be placed either in front of or after the indirect object: *P •& W f % (M98Af|7/r*b '... and sent an announcement of his difficulties to Ts'e.' (Legge, p.150) $ % j&- f -f (HY 278/^.14/17x) 'Woo announced to Tsin the defeat ' (Legge, p.463) 'and how he repented of his oath.' (Legge, p.6) 'and (the duke) told (Ying K'ao Shu \\ about his repentence.' (the author's translation) ^ *. * <* «? 1 <SI 'and he would tell him how to distribute the bribes.' (Legge, p.698) 'I am going to tell you how to distribute the bribes.' (the author's .translation) These structures appear to be parallel to the structure 'ff/^.-J ^fcfc $L yu-offer a captive to Ancestral Mother Keng' ,(Ping 47 (21)) and 'ff f £ % ^ yu-offer to Wang Hai a female slave' (Ping 117 (30)) respectively. And it - 199 -seems to be possible to treat the O.B.I. R-S verb kao ^ , for instance, in a similar way: + OG + C J ) + OB e.g. j£ ^ (Ts'ui ffiJE 1107) We should announce (the activities) of the T'u Statelet to Shang Chia.' •It + ( ) + OB + 0V e.g. ^ J J | ^ (Nan Ming Jfr 616) We should make an announcement to Wang Hai (with) five oxen. Or: We should announce-sacrifice (i.e. the sacrifice incorporated in the ritual 'announcing') to Wang Hai five oxen. However, there is solid evidence showing that the difference between an OG and an OV is not only semantic but also syntactical. That is, within the patterns 'VA + ( ) + OB + 0' and 'VA + 0 + ( ^ ) + OB', an 0V almost invariably appears in the former while an OG appears in the latter, i.e., 'V + ( ^ ) + OB + OV' and 'V + OG + ( ^ ) + OB'. Following is a brief statistics (data are collected from the Sorui with the partly obliterated and obscure inscriptions excluded): V + OV + ( 5" ) + OB occurrence: Q V + ( "j" ) + OB + OV occurrence: 18 V + OV + ( -j" ) + OB occurrence: 1 V + ( -j£ ) + OB + OV occurrence: 20 - 200 -6 V + OV + ( K ) + OB occurrence: 1 V +•(. ^ ) + OB + OV occurrence: 12 V + OG + ( j" ) + OB occurrence: 38 V + (. f ) + OB + OG occurrence: l7 1 V + OG + ( -f ) + OB occurrence: 64 V + ( ~f ) + OB + OG occurrence: 0 V + OG + ( T ) + OB occurrence: 104 V + ( j ) + OB + OG i8 occurrence: 1 The OGs and OVs can co-occur in one single inscription, e.g. perhaps / announce / locust / Shang Chia / two / ox Ts'ui 88 We should perhaps announce (that there is) locust to Shang Chia (with) two oxen. I ... hsU / crack / ... / exorcise / Prince ft (?) / to Mother Chi / two / small / penned-sheep We should exorcise Prince ^ (?) (the exorcism being directed) to Mother Chi (with)two small penned-sheep. Nan Wu 134 test / invoke'/ harvest / to / T'ai Chia / ten / penned-sheep Ancestor I / ten / penned-sheep Ping 117 (24) - 201 -We should invoke for a (good) harvest to T'ai Chia with ten penned-sheep, and to Ancestor I faith) ten penned-sheep. (Notice that the order of the OGs and OVs is maintained, i.e., the OGs are placed immediately after the verbs while the OVs are separated from the verb by the OBs.) Now it appears that type A verbs are four-place verbs; they can potentially take an agent (normally unexpressed), an OB, and 0G and an OV. By referring to the' classics, we may reveal the true relationship between the type A verbs and the OVs. M f£ £^Mt -? % (Tso Chuan HY 82/^2/5^. ) '... sent bribes to Keu, and requested the delivery of Kung-chung.' (Legge, p.129) '... used bribes to request the delivery of Kung-chung from Chtt.' (the author's translation) (Tso Chuan HY 239/ ^16/ jffil) '... the viscount of Ts'oo sent the Kung-tsze Ch'ing from Woo-shing to seek for peace with Ch'ing by the offer of the lands of Joo-yin.' (Legge, p.395) The objects 'bribe' and ^® 'field' are the things used, as a means or instrument, to obtain the goal sought. They can be termed 01s (obj ect-instrumeht) introduced by the coverb 1^ . ^ - 202 -The word 1 vjA , in classical Chinese functions as a co-verb and still maintains a strong verbal sense. Cf. ^ V>A *\% )jt\ ^(Tso  Chuan HY 2/5"^ ) *Ch'ing-ke, immediately on the duke's death, had gone to Choo, taking with him duke Chwang's remaining son, who was afterwards duke He' (Legge, p.129). We may even put it this way: in relation to the beneficiaries, i.e., ChU ^ and Cheng jjjPj) , there are two verbs; one is the word ch'iu pft. which takes the OG, the other is the word 1_ which takes the 01 and these two verbs both take the beneficiary /J^p as an indirect object. ^ It is reasonable to assume that the sentence ' v\ S^L l"t "f ' > for instance, is derived from a hypothetical primitive structure '*to take/use a bribe to request the delivery of Kung-chung from ChU.' Cf. Legge uses a full verb 'sent' to translate the word If the above assumption is justified, we may make attempt to analyse the O.B. sentence in a similar way: £ ^ J < ^ ^ Jjjp (cited on p.200) We should invoke for a (good) harvest to T'ai Chia (by USING) ten penned-sheep. Or: In praying for a (good) harvest to T'ai Chia, we should USE ten penned-sheep. In the underlying structure, there are two verbs, one takes the OG, ; the other takes the 01, i.e., These two verbs, on the other hand, share a common 0B, i.e.,-7X^ - 203 -This hypothesis has the merit of explaining the following phenomenon exemplified by the verb ch'iu z£o • The following structures frequently appear: ;"f<L + OG + (. ) + OB + OV e.g. (3*7 J £ ^ Ping. H7 (25) We should invoke for rain to T'ai Chia (with) a penned-ox. + OB + OV A + ( e.g. j£ ^ /^V' f T ch,1en ty) 1.44.2 We should invoke to Hsien (with) ten oxeri. h + OG + ( ^ ) + OB, VB + OV, (VB + OV) In praying for a (good) harvest to the Mountain(god), we should burn-sacrifice three small penned-sheep and dismember-sacrifice three oxen. Ho_ 340 ( ) + OB, Vfi + OV, (VB + OV) e.g. ^ ff^ >% ^ In praying to T'ai Chia, we should burn-sacrifice one penned-sheep, three dogs and dismember ... Ling But the following structures do not appear: * T[L + OG + ( "f ) + OB + OV, VB + OV, CV„ + OV) * + I "f ) + OB + ov, vB + ov, (vB + ov)11 - 204 -In other words, wherever there is a structure 'V + OV' (e.g. 'a. 4" '% ), we do not find an OV following right after the structure ' jj< + (OG) + ( -j- ) + OB'. By assuming there is an unexpressed abstract verb (comparable to the coverb JL >|A in classical Chinese) , we can provide the following hypothesis: In the structure ' ^ + (OG) + (_ f ) + OB + OV', there are in fact two verbs (or even three, if we count the co-verb yjl -j~ ), i.e., ' rfz. + (OG) + ( ) + OB, USE + OV'. The verb USE takes the OV and'may be specifically realized, e.g., tfc + (OG) + ( ^ ) + OB, ^ + OV'. No doubt, the hypothetical structure '* + (OG) + ( J- ) + OB + OV, Vfi + OV' is not attested because it is clumsy and redundant, i.e., '*in praying for (X) to so and so, we should USE a Y, we should burn-sacrifice a Y '(where 'burn-sacrifice' is interpreted as the realization of the abstract verb USE and the two Ys are interpreted as referring to the same victim). The above hypothesis can also account for the fact that the pattern ' + ( ^ ) + OB + OV' occurs so frequently but the pattern ' ^. + OV + ( *f ) + OB' almost never occurs. The relationship between the verb ch'iu ^ and the OB is much closer than that between ch'iu and the 0V(=0I). In fact, an abstract verb USE may even be said to separate the main verb and its OB from the OV. Sometimes, the abstract verb USE is expressed in the surface structure without being semantically characterized, e.g., to / Brother Ting / exorcise / use / ox When performing an exorcism to Brother Ting, we should use an ox. I. f\'%^ 449 - 205 -Unlike the structure where the methods of sacrificing the victims are specified, e.g. ^ burn-sacrifice, dismember-sacrifice, the above structure carries an information load almost identical to that of * -J" ^ %f (or fyf "f ^ ^ ty That is' we sti11 do not know now this R_s activity is going to be performed. Probably for this reason, this structure and the structure ' tyf + f + OB, f£) + OV' are rarely found in the O.B.I. Now let us look at the verbs B4, B5, B6 (liao ^ , vja and vu_ ^). They obviously differ from the verbs Al, A2 and A3 in the following aspects: (1) They do not take an OG. (2) The OV enjoys a rather flexible position, i.e., the patterns 'VB+ OV + ( ^ ) + OB' and 'Vfi+ ( ^ ) + OB + OV' 12 are both attested in the O.B.I. (3) The pattern 'V_ + ( -f ) + OB, VD + OV, (Vn + OV)' does not B B . s-> occur; instead, we find the pattern 'Vg + ( ^ ) + OB + OV, vB + OV, (vB + OV)'. For example, f13 B«^^/8, \ V ft$ * % ff kuei-yu / crack / test / burn-sacrifice / to Mountain(god) / three / small / penned-sheep j> <vt =• s. W if dismember-sacrifice / three / penned-sheep We should burn-sacrifice to the Mountain(god) three small penned-sheep and dismember-sacrfice three penned-sheep. Ch'ien ff) 7.25.3 - 206 -B4u U fl*. 1 > ? n f vU jen-ch'en / crack / next / chia-wu / burn-sacrifice to / tt O) / sheep / yu-offer / pig On the next chia-wu day, we should burn-sacrifice to (the spirit of) \\ (?) a sheep and yu-offer a pig. Ch'ien 4.52.4 BMiife. ft U V f N § * * 'i-ssu / crack / Cheng / test / burn-sacrifice to / River(god) / five / ox / sink-sacrifice / ten ox / ten month / being at / Tou(?) We should burn-sacrifice to the River(god) five oxen, sink-sacrifice ten oxen. (Divined at) the tenth month, in Tou(?). Ch'ien ft) 4.52.4 •it B5i ,{iz ''9*s # J z. test / come / i-hai / yu-sacrifice / Ancestor I | -r ^ la ^ *p It ten / behead / and / five / dismember-sacrifice / ten penned-sheep I_ Z- 6408 On the coming i-hai day, we should yu-sacrifice to Ancestor I fifteen beheaded victims and dismember-sacrif ice ten penned-sheep. i-mab / crack / Ch'ueh / test / come / i-hai yu-sacrifice / Hsia I / ten / behead / and / five f I -r ¥ ^ - ^ dismember-sacrif ice / ten / pennedr-sheep / two - 207 -ten-day-week / and / one / day / i-hai not / yu-sacrif ice / rain Ping 197 (.3) On the coming i-hai day, we should yu-sacrifice to Hsia I fifteen beheaded victims and dismember-sacrifice ten penned-sheep. (After) twenty one days, on an i-hai day, we did not perform the yu-sacrifice. It rained. tf156i H§ fi g# Y f •••• a IN test / yu-offer / to Ancestor Hsin / five / behead dismember-sacrifice / three / penned-sheep Ping 27 (1) We should yu-offer to Ancestor Hsin five beheaded victims and dismember-sacrifice three penned-sheep. 6ii £3 £ ¥ *a ) / + test / yu-offer / Ancestor I / ten / behead dismember / three / ox Ping 102 (7) We should yu-offer to Ancestor I a beheaded victim and dismember-sacrifice three oxen. Although the structure 'V + OV' (e.g. ) appears, there is still an OV following right after the structure 'V + ( ^ ) + OB' (e.g.-^. > ^). Such a structure excludes the possibility that type B verbs do not take an OV which is governed by a 'specific realizing' verb. It appears to be a very remote possibility that the Shang would, frequently and regularly, make divinations in the following way, 'in performing a burn-sacrifice to the Mountain(god), we should USE (method unspecified) three small penned-sheep and dismember-sacrifice (method specified) three penned-sheep' (Ch'ien 208 -7.25.3). There is a vast difference in the semantic structure between the type A and type B verbs although they both appear in the pattern 'V + (. -j" ) + OB + OV'. The relationship between the type A verbs and the OVs are rather loose, it takes another verb (i.e. USE) to conjoin them. On the other hand, the type B verbs really 'take' the OVs as patients and they represent a specific method of sacrificing victims. Cf. the OVs can be placed between the type B verbs and the OBs, i.e. 'V + OV + C })+ OB'. B The following verbs show linguistic behaviour virtually similar to that of liao J| , yju and v_u htl , therefore, they are also classified as type B verbs: 7. "(jjf 'announce-sacrifice' (Li Hsiao-ting, 1965:1606) 8. ( ^"f ) 'chop-sacrifice' (Yd Hsing-wu, 1941:29) 9. '/-Q- 'tsu-sacrifice (to present victims on a tsu-stand) ' (Li Hsiao-ting, 1965:4089) (A) These verbs never appear in the pattern ' ^ / $ / K'B- + ( -f ) + OB, vB + OV, (vB + OV)' On the contrary, they appear in the pattern ' / tfj? /'/$*+( ^ ) + OB + OV, V^ + OV, (V + OV)' B B For example, $ 7i n g s # *\ *ti \f test / announce-sacrifice / Ancestral Mother Keng / ten %«. i <*? i + f % captives / dismember-sacrif ice / ten / penned-sheep I_ "21 751 We should announce-sacrifice to Ancestral Mother Keng ten captives and dismember-sacrifice ten penned-sheep. - 209 -711 9i test / announce-sacrifice / Ancestor I / ten / behead V*. Xi f «r It "^t and / five / dismember-sacrifice / ten / penned-sheep and / five Ho ^ 254 We should announce-sacrifice to Ancestor I fifteen beheaded (victims) and dismember-sacrifice fifteen penned-sheep. H# f=f ftte&d ff test / call upon / Lady / chop-sacrifice / to / Father I Wi & IF • • = * penned-sheep / announce-sacrifice / three / penned-sheep yu-offer / captive - Ping 182 (11) We should call upon the lady to chop-sacrifice to Father I a penned sheep, announce-sacrifice three penned-sheep and yu-offer a captive. kuei-yu / tsu-sacrifice / to / ^ (?) / Ch'iang / three U IT ' It ty* ^ people / dismember-sacrifice / ten / ox / right (?) We should tsu-sacrifice (i.e. to present the victims on a tsu-stand) to ^ (?) three Ch'iang-tribesmen and dismember-sacrif ice ten oxen. Right(?). HsU 1.52.2 9ii {LL%%. Mm ff $L f chi-wei / tsu-sacrifice / to / «f(?) / Ch'iar dismember-sacrifice / ten / ox / middle (?) We should tsu-sacrifice to ^ (?) three Ch'iang-tribesmen 16 and dismember-sacrifice ten oxen. Middle(?). Ch'ien 6.2.3 - 210 -These three verbs appear in patterns highly similar to those of the verbs liao % , y_u ^(31 and y_u ^ . Although the structure 'V,, + OV' (e.g. + ^ ) appears, one still finds an OV following the structure ' 3 / $ I '& + ( f ) + OB', i.e., jg / £ /£B + ( f ) + OB + OV. This* fact suggests that these three verbs also represent specific methods of sacrificing victims. (B) These three verbs appear in positions parallel to that of liao ^ , yu y_u -y when there is a preceding type A verb, for example, chi-mao /crack / Ch'Ueh / test / exorcise / (Lady)Hao ft 4. ft t1* f ?< to / Father I / chop-sacrifice / sheep / yu-offer / pig announce-sacrifice / ten / penned-sheep I_ 3383 In exorcising Lady Hao (the exorcism being directed) to Father I, we should chop-sacrifice a sheep, yu-offer a pig and announce-sacrifice ten penned-sheep. hsin-... / test / invoke I ... I River(god) / burn-sacrifice five / small / penned-sheep / sink-sacrifice / five / ox dismember-sacrifice / five / ox / tsu-sacrifice / penned-sheep In invoking for ... to the River(god), we should burn-sacrifice five small penned-sheep, sink-sacrifice five oxen, dismember-sacrif ice five oxen and tsu-sacrifice a penned sheep. Ho ^ 76 - 211 -crack / Cheng I ... I Prince"^" (?) / to ft % W$ mother / chop-sacrifice / pig / small / penned-sheep yu-offer / captive / woman / one Ping 512 (1) In (exorcising(?)) Prince"^- (?) (the exorcism(?) being directed) to Mother (Ping), we should chop-sacrifice a pig, a small penned-sheep and yu-offer one female captive. ('Mother Ping' is an ancestral name appearing on the same plastron. Probably the word in the above inscription also refers to Mother Ping. The word 'exorcise' has been supplied here on the basis of a comparison with other similar inscriptions, e.8. H ft T t > ' S ^ ••• (Ching % 2088) 'In ex-orcising Prince YU (the exorcism being directed) to Father I, we should chop-sacrifice a sheep ...') Mi fevf 1tar test / exorcise / Prince / Pin / to / Brother Ting chop-sacrifice / sheep / announce-sacrifice / small penned-sheep / this / day / yu-sacrifice Pu ' 288 In exorcising Prince Pin to Brother Ting, we should chop-sacrifice a sheep, announce-sacrifice a small penned-sheep; and, today, perform a yu-sacrifice. -212 -Obviously, these verbs can also be interpreted as type B verbs representing specific methods of sacrificing victims to realize or accompany the type A R-S activities. 2. The Relationship between the Type A and Type B Verbs After illustrating the difference between the type A and type B .verbs, we may proceed to investigate the relationship between them. jen-tzu / test 7 perhaps / invoke / wheat / to River(god) / burn-sacrifice / three / penned-sheep HI «t H * | & ^ $ sink-sacrifice / three / tsu-sacrifice / penned-ox To_ ipf 2.404 If we perhaps invoke for wheat to the.River(god), we should burn-sacrifice three penned-sheep, sink-sacrifice three oxen and tsu-sacrif ice a penned ox.^ i t ^ ^* ff chia-tzu / crack / Cheng / test / invoke / harvest / to K*uei / burn-sacrifice / six / ox Chia 7 3512 In invoking for a (good) harvest to K'uei, we should burn-sacrifice six oxen. In exorcising Lady Hao (the exorcism being directed) to Father I, we should chop-sacrifice a sheep, yu-offer a pig and announce-sacrifice ten penned-sheep. (Cited on p. 210.) . - 213 -test / exorcise / Prince YU / to / Father / yu-offer one / behead / dismember-sacrif ice / penned-sheep Ching' /jt, 807 In exorcising Prince YU (the exorcism being directed) to Father (X), we should yu-offer a beheaded victim and dismember-sacrifice a penned-sheep. X* -:/U H JS ^ £ * i£*p $4 kuei-ch'ou / test / perhaps / big / exorcise / hui chia-tzu / yu-sacrifice Ts 'ui "4^ 79 If we will perhaps perform a big exorcism, it should be on a chia—tzu day that we perform a yu-sacrifice. wei / test / perhaps / exorcise / Kung / to / ,... ping-shen »h// Y A i* > 1 yu-sacrifice / fifty /small / penned-sheep Ts'ui Tj 502 If we will perhaps exorcise Kung (the exorcism being directed) to we should, on the ping-shen day, yu-sacrifice fifty small penned-sheep(?). U t< Hi % i f J U °T chia-tzu / crack I la I test / announce / to / Father Ting hui / this / day / wen-sacrifice (?) / yu-sacrifice I_ 7^ 881 If we are going to make an announcement to Father Ting, it should be today that we perform a wen-sacrifice(?) and a ... 18 yu-sacrxfice. As shown in the translations, we interpret the type B verbs as words representing minor R-S activities which accompany the major R-S activities - 214 -represented by type A verbs. In other words, the above inscriptions are divinations performed to determine what specific sacrifice (type B verbs) should be carried out as realizations of or in accompaniment to type A R-S activities. Type B R-S activities are subordinate to type A R-S activities. But it should be noted that in syntactic analysis, type B verbs constitute main clauses since they are the focal point, i.e., the activities to be determined by the divination, while type A verbs constitute conditional or adverbial clauses. This interpretation seems to be the most natural one as it can explain the different linguistic behaviour of the two types of verbs. The spirits function as the beneficiaries of type A verbs, e.g. ( ) one who is being invoked* and also as beneficiaries of the sacrifices, e.g. ( jjj^ ) the one who receives the victims used in a burn-sacrifice. Therefore, we find both structures like ' ^ ^ i°J ' to invoke for a (good) harvest to the River (god) (jD^^, 375) and structures like ' jjl 1®J ">%$\j to burn-sacrifice to the River(god) three boars) (Jin A< 2985). When structures similar to the above two are combined in a single inscription, normally the OB of the type B verb is omitted since the two OBs are co-referential and a structure similar to Chia ^ 3915 occurs. (In invoking for a (good) harvest to K'uei, we should burn-sacrifice six oxen.) - 215 -Alternatively, type A and type B verbs may appear in independent, but related, divinations. Thanks to Chang Ping-ch'Uan's efforts in reconstructing fragments into complete plastrons, we are equipped with a larger context from which we may make a reasonable hypothesis concerning the sequence of divination. Ping 32 and 33 (the two sides of one plastron) serve as a good example. test / king / misfortune / favorable Ping 32 (17) The king (is experiencing) misfortune, he will be all right. Father I /. harm / king Ping 32 (15) Father I harmed the king. Or: Father I is harming the king. ^ #* ft ti Father I / not / harm / king Ping 32 (16) Father I did not harm the king. Or: Father I is not harming the king. Ma f*- * B 9 } + test / not / fortunate / announce-sacrifice / ten Ancestor I Ping 32 (7) It was not fortunate. We should announce-sacrifice ten 19 (victims) to Ancestor I. should not / to / Ancestor I 33 (4) We should not (perform a sacrifice) to Ancestor I. - 216 -^ )h li ft contribute / Father I / ten / Ch'iang 33 (9) We should contribute to Father I ten Ch1iang-tribesmen. (See T'ang Lan 1939:47 for the interpretation of ^( ^ — f# ) .) should not / contribute-sacrifice 33 (10) We should not contribute-sacrifice (to Father I). kuei-hai / crack / Ch'Ueh / test / exorcise / to Ancestor Ting 32 (1) We should perform an exorcism to Ancestor Ting. St ° T I t "ft $ announce-sacrifice / Ancestor Ting / ten / behead ten / penned-sheep 32 (2) We should announce-sacrifice to Ancestor Ting ten beheaded-victims and ten penned-sheep. should not / announce-sacrifice / Ancestor Ting We should not announce-sacrifice to Ancestor Ting. 32 (3) Ping 32 (17) ' X- S5 ' shows that some misfortune has befallen the king. And Pina 32 (.15), (16) ' '£ L % /$ 2- ' ; 32 (7) ' f^-f- L ' , 33 (4) ' /TJf f 4fl_ L''> 32 -CD ' f %L J"' show that in the mind of the 20 Shang, there are three ancestors who may be causing the king's misfortune. Ping 32 (.1) divines about whether they should perform an exorcism to Ancestor Ting. (There is no negative counterpart '* //7J ^tf -f %SL J ' •) Presumably, the 'response' of ' 4~f f /^JsL X ' was positive, so the Shang went on to divine - 217 -about what kind of sacrifice, and what sacrificial-victims should be used, i.e., 32 (2) T + 4% \ % and 32 (3) 4$ %B- T > That the announce-sacrifice ( "0" ) is used as a specific realizing verb to carry out an exorcism (. JfJI ) can be illustrated by the following plastron which also serves an an example to show the relationship between type A and type B verbs: .^\t-.Vf L-i-:tf tff& test / Lady Hao / perhaps / recover / ju / sickness Ping 549 (1) 21 Lady Hao will perhaps recover from sickness. (Yu ^ is interpreted as an honorific marker, comparable to the expression kuei yang ^ JZ. , a proposal made by Takashima in one of his seminars, and subsequently published in Takashima 1980a:63.) .. . Hao / not I ... I recover / ^yju .. . (Lady) Hao will not ... recover ... 549 (2) test / exorcise / Lady Hao / to / Father I / chop-sacrifice 'ft * • S' e. S # it penned-sheep / yu-offer / piglet / announce-sacrifice / ten penned-sheep / ten / captive / piglet / ten 548 (3) In exorcising Lady Hao (the exorcism being directed) to Father I, we should chop-sacrifice a penned-sheep, yu-offer a piglet, announce-sacrifice ten penned-sheep, ten captives, and piglets amounting to ten. - 218 -should not / announce-sacrifice / Father I / ten *<U It (ft k& it captive / ten / penned-sheep / piglet / ten 548 (4) We should not announce-sacrifice to Father I ten captives, ten penned-sheep and piglets amounting to ten. keng ... / chop-sacrifice / penned-sheep / yu-offer / penned-sheep i + piglet / ten 548 (1) On the keng ... day, we should chop-sacrifice a penned-sheep, yu-offer a penned-sheep ... and ten piglets. Ping. 549 (1), (2) • f ' show that Lady Hao is sick and the Shang divine about what R-S activities should be performed to Father I. The $ and ^ , as a set, seem to be one alternative, and $ seems to be another alternative; they are divined about separately in Ping 548 (1) and (4) respectively. Or both alternatives may be practiced, as shown in 548 (3). .Ping 548 (3) and (4) nicely exemplify the relationship between type A and type B verbs. In both inscriptions, Father I *?£ ^ functions as the OB. But in 548 (3), Father I appears as the OB of the verb ty^ while in 548 (4), the OB of the verb ^ . The fact that both and 4^ share a common OB and if^7 does not take a object-victim where one finds the structure 'V^ + OV', i.e., ' ^ + f + (and also ' $ + f ', ' ^ + ') » corroborates the theory that type B verbs embody specific options for the realization of type A verbs. I - 219 -3. The Classification of Some Other Verbs In the case of the six verbs cited above, i.e., kao , ^u^ » ch' iu , liao , fyfy and yu_ ^ , regular and neat patterns are revealed. However, such patterns are not always exemplified by other verbs. Sometime we have to rely on other means in making a classification. For example, ) 'dismember-sacrifice' (Wang Kuo-wei, 1917:5-6) b. y^U 'sink-sacrifice' (Lo Chen-yU, 1914: Vol. 2, p.16) c. j|j ( !?]fj ) 'split-sacrifice' (Yti Hsing-wu, 1941:19-20) - J>) V These verbs do not appear either in the pattern ( ^) + OB, Vg + OV, (Vg + OV)' or the pattern ' / »%> / jfjj + ( ^ ) + OB + OV, Vg + OV, (V + OV)'. Nevertheless, they appear in positions parallel to other type B verbs: a. 'sink-sacrifice' hsin-wei / test / invoke / grain / to / River(god) / burn-sacrifice three / penned-ox / sink-sacrifice / three / ox / tsu-sacrifice penned-ox In invoking for grain to the River(god), we should burn-sacrifice three penned-oxen, sink-sacrifice three oxen and tsu-sacrifice a penned-ox. Ning ig; 1.119 J - 220 -2en-shen / test / invoke / grain / to / River(god) burn-sacrifice 7 three / ox / sink-sacrifice / three ox Jin A 2361 In invoking for grain to the River(god), we should burn-sacrifice three oxen and sink-sacrifice three oxen. ) 'dismember-sacrifice' Hit * * f? ii = * -<-^ test / burn-sacrifice I to I Earth(god) / three / small penned-sheep / dismember-sacrifice / two / ox / sink-sacrifice ten / ox Ch'ien ty) 1.24.3 test / invoke / harvest / to / Mountain(god) / burn-sacrifice 3 .'• * ^ jf f *J> = r three / small / penned-sheep / dismember-sacrifice / three ox / two /rmonth In invoking for a (good) harvest to the Mountain(god), we should burn-sacrifice three small penned-sheep and dismember-sacrifice three oxen. At the second month. Ho ^ 340 W C ff>'j# ) 'split-sacrifice' ••• mao / vu-offer / to / Mother Hsin / three / penned-sheep split-sacrifice / one / ox / Ch'iang / ten ... Chin 694 - 221 -... map day, we should yu-offer to Mother Hsin three penned-sheep, split-sacrifice three oxen and ten Ch'iang-tribesmen... ... wu / crack / Ta / test / next / kuei-wei / yu-offer f i A ... 3\ = i. f t w$-to / minor / <K. (?) / three / penned-sheep / split-sacrifice -one / ox On the next kuei-wei day, we should yu-offer to the minor (?) three penned-sheep and split-sacrif ice one ox. HsU %g 2.18.1 The verbs map and ch'en vj£, appear in, positions, parallel to the word liao JfL in inscriptions Ning i^,1'. I'jl9, Jin X^.2361 and Ho 340, we can interpret the former as type B verbs. In inscriptions Chin 694 and HsU 2.18.1, although the type B verb yu takes an OB while fu&'Jjp (Jjj) does not, we may still interpret these two verbs as coordinate. When the first verb in a sequence of two or more coordinative type B verbs takes an OB, the following structure is usually found: 'VB + ( "f )- + °B + °V' VB + °V' (S6e Page 205~207)-Compare the inscriptions Ho 340 and Ch' ien fli] 1.24.3: Because of the parallel and/or coordinate relationship between the verbs mao ^ ), ch'en V^L , fu&'|<|? C 1$• ) and the attested type B verbs liao ^ and ;y_u ^ , the former three verbs are also classified as type B. - 222. -Furthermore, if we modify our original notion of the difference between type A and type B R-S activities (I.e. type A refers to R-S activi ties the performance of which requires other R-S activities, and type B refers to R-S activities used to accompany other R-S activities), we may extend the categories A and B to verbs which do not take OVs. The word kuan , as mentioned in footnote no. 1 of this chapter, rarely takes an OV. (The OV ( ) is incorporated in the semantic structure of 'ffifi (to pour wine).) Therefore, it lacks one of the character istics of type B verbs mentioned above. Nevertheless, in the following inscription, kuan shows an apparent coordinate relationship with the type B verbs ^u fj^ and sui both of which can be interpreted as R-S activities accompanying the type A R-S activity kao "£ . (See p.250 for jt. the interpretation of sui ^ .) ting-ssu / crack / Ch'Ueh / test / announce / misfortune to / Ancestor (I) / should not / yu-offer / sui-cut-sacrifice P 'if perform libation In announcing the king's misfortune to Ancestor (I), we should not perform yu-offering, sui-cut-sacrifice and libation. Ping 98 (10) In some other inscriptions, the verb:kuan Tfffi shows an contrastive relationship with either the verb sui ^ C Jl'J ) or ju_ ^ , as in: - 223 -( A/ Jk \ , is/1 )* G kuei-mao / crack / test / hui / perform libation <j It should be a libation (that we perform). Chia ^ 3915 kuei-mao / crack / test / hui / sui-cut-sacrifice It should be a sui-cut-sacrifice (that we perform). Chia T" 3915 We should perform a libation to Ancestor I in order to announce the king's misfortune. Ping 98 (12) We should yu-offer to Ancestor I in order to announce the king's misfortune. Ping 98 (14) (For a word-for-word translation of the last two inscriptions see p. 226). Whether these verbs appear in a coordinative or a contrastive relationship, the same interpretation can be maintained, i.e., these verbs are of the same 'rank' in relation to type A verbs such as kao . The performance of a libation to accompany an announcement seems to have been a long-lived practice in Chinese religious culture. The phrase found in K'ung An-kuo's l|$ commentary to the Shang Shu ' ^ $ & to pour aromatic wine to announce to the spirits' (SSCCS, chUan 15, p.27; cited by Shima 1958:268) matches these inscriptions nicely. In the Li Yun Chapter of the Li Chi, there is a passage about libations and announcements in a sacrificial context. - 224 -(SSCCS, Chuan 21, p.16) To use hsUan-wine (water) to perform rituals, (then) present the blood and hair (of the victims) ... in the chu-praying, one announces the filial piety; in the chia-22 praying, one announces the patriarchal kindness.' Thus, it seems justifiable to classify kuan as a type B verb even though it does not show the characteristics found with other type B verbs. On the other hand, verbs such as i ( ^ ) or i-jih (one of the five sacrifices performed according to a regular schedule) do not take an OG or an OV. Therefore, the patterns ' vf\\ + ( 3$ ) + OB, Vg + OV, (V_ + OV)' and ' fit] + ( 3* ) + OB + OV, V_ + OV, (VB + OV)' we have used a = B B as one of the criteria to distinguish type A verbs from type B verbs are not applicable. However, since these R-S activities are performed according to a regular schedule, presumably, they are major R-S activities. And we may interpret the type B verbs which co-occur with them as R-S activities used to accompany these major R-S activities. For example, chia-tzu / crack / i-j ih / Ju I / yu-offer / ascend-sacrifice sui-cut-sacrif ice / three / ox I <L 5394 In performing an i-j ih ritual to Ju I, we should yuroffer, ascend-sacrifice and sui-cut-sacrifice three oxen. (The author follows Ch'en Meng-chia (1956:417) in identifying A j as -J' "Z. Hsia I.) - 225 -kuei-ssu / crack /^ C?) / test / i-jih / Ancestor Chia sui-cut-sacrifice / perhaps / penned-sheep Ts'ui ^-f 333 In performing an i-jih sacrifice to Ancestor Chia, we should sui-cut-sacrifice; perhaps a penned-sheep. I - 226 -II. ANALYSIS OF THE STRUCTURE.'... V_ ... B A In the previous section, we have investigated the relationship and the contrast between type A and type B verbs. In the majority of the inscriptions cited above, type B verbs are placed after the type A verb and we provide the analysis 'if we do (or, in doing) A, we should do B (to realize or to accompany A)'. In contrast to the above structure, there are cases where type B verbs are placed in front of type A verbs. We will attempt to investigate and interpret this structure in the present section. CD Ping 98 ting-ssu / crack / Pin / test / perform libation / to Ancestor I / announce / king / misfortune Ping 98 (12) We should perform a libation to Ancestor I in order to announce the king's misfortune. test / should not / specifically / perform libation / to AM ^ %z I *(& Ancestor I / announce / misfortune 98 (13) We should not specifically perform a libation to Ancestor I in order to announce the king's misfortune. yu-offer / Ancestor I / announce / king / misfortune We should yu-offer to Ancestor I in order to announce the king's misfortune. 98 (14) - 227 -should not / yu-offer / perform libation / Ancestor I 98 (15) We should not yu-offer and perform a libation to Ancestor I. ting-ssu / crack / Ch'Ueh / test / announce / misfortune 1* i M W /c to t1 * fa ft to / Ancestor (I) / should not / yu-offer / sui-cut-sacrifice perform libation 98 (10) In making an announcement of the king's misfortune to Ancestor I, we should not perform a yu-offering, a sui-cut-sacrifice and a libation. test / king / not / perform libation / tablet / obstruct 98 (16) If the king does not perform a libation, the (spirits of the) tablets will obstruct. test 7 tablet / will not / obstruct / king / not / perform libation 98 (17) The (spirits of the) tablets will not obstruct (because of) the king's not performing a libation. All these inscriptions appear on the same plastron and obviously are related to one issue — what R-S activities should be performed to announce the king's misfortune to Ancestor I. There are three R-S activities proposed-yu 4i , sui 7^/ and kuan . As in the case of most plastrons, we do not find regular and parallel patterns. The sui sacrifice is mentioned only once (98 (10)) while the kuan^>j^ ritual is mentioned five times, twice in coordination with sui and yu ^ . It is illuminating to compare the - 228 -inscriptions 98.(10) and 98 Q-2), (.13). In 98 (.10), the type A. verb kao is placed in front of the type B verb kuan %M- while in 98 (12), (13) show Ah the opposite arrangement. Syntactically, 98 (12) can be interpreted as '*in performing a libation to Ancestor I, we should announce the king's misfortune.' However, Ping 98 (13) where the negative wu ^JJ is placed in front of both verbs (i.e., 'T/J ... + /fjj[+ ... % ...) excludes the conditional analysis proposed above. The cumulative interpretation, i.e., 'we should/should not perform a libation and announce the king's misfortune', is also unlikely, since in the inscription.' % /[ffi f /fj[ ( ) , 'iT] .^f Jjj; fiffi the negative wu is placed in front of the verbs yu_ , sui and kuan pjjL rather than in front of all the verbs, i.e.,'* #7 ^ J0j f /fg_ ZJ » ^ The most natural analysis appropriate to the larger context is 'we should/ should not perform a libation to Ancestor I to announce the king's mis fortune' as proposed in the translations. The R-S activity kuan Jsffi seems to have enjoyed a particular significance. With the exception of Ping 98 (14), it appears in all six other inscriptions. In Ping 98 (16), (17), the Shang explicitly divine about the consequences of the performance (or more precisely, the suspension of) the libation. In fact, these two inscriptions show a rather uncommon pattern. In the first place, the clause ' % rfi fpjf: ' is shifted from the initial position to the final position in Ping 98 (16) ' £ S)\ p$r , jfs -f ', the negative p_u ^x is placed in the protasis. As mentioned in Chapter III, p.122, the negative sense in a conditional or simultaneous-successive sen tence is normally expressed in the apodosis, i.e., '* % » % ~f r^y t. 7[- 7)- j~ ' (.*if. the king performs a libation, the (spirits of the) tablets will - 229 -obstruct if the king performs a libation, the (spirits of the) tablets will not obstruct'). As observed in this inscription, such pre-positioning may be due to 'emphasis' (as Takashima has suggested, 1973:80. The above translations have been suggested by Takashima, although the author interprets the graph somewhat differently from Takashima). Also notice that the negative of kuan jf^ is p_u rf, instead of the wu found in Ping 98 (13), (15) and (10). In other words, the sense of 'controllability' has been neutralized as may be inferred by the shift of the negative wu ^7 to p_u rf- . In addition to the interpretations reflected in the translations, there seems to be another alternative. When these two divinations were made, the Shang had already decided not to perform the libation and indeed they did not, since the ritual process had already reached the kao ^ stage and the appropriate time for the performance of a libation had passed. The Shang, for reasons we can hardly speculate on, may have been concerned about the suspension of the libation and thus, they divined about the consequence of this suspension. If the above assumption is correct, the clause ' ct rfK ft^ 1 is simply a statement (record) of an activity (or non-activity) in the past. As it is a past activity, the question of controllability does not arise, therefore, the modally unmarked negative p_u rfx is employed. Adopting this interpretation, we may translate these two inscriptions in the following way: Ping 98 (16) 'The king did riot perform a libation, the (spirits of the) tablets will obstruct.' Or: 'As to the king's not performing a libation, (the spirits of) the tablets will obstruct.' - 230 -Ping 98.(17) 'The (spirits of the) tablets will not obstruct (because of) the king's not performing a libation.' Unfortunately, inscriptions showing the above patterns are extremely rare (the above two are the only occurrences in Ping Pien), so it is very difficult to substantiate this alternative interpretation. (2) Ping 114 test / yu-offer / to / (Shang)Chia / three / penned-sheep ounce / we / Pao / shackle announce We should yu-offer to (Shang)Chia three penned-sheep to announce (the military success(?)) to the Pao spirits (Pao I Z, Pao Ping^g (fa , Pao Ting J ) with shackled 23 (captives). 114 (8) test / one / penned-sheep I to I (Shang)Chia / announce we / Pao / shackle^ 114 (9)' (It is) one penned-sheep that we should yu-offer to (Shang) Chia to announce the (military success(?)) to the Pao spirits (Pao 1^1/, Pao Ping 2jf/£ , Pao Ting J ) with shackled (captives). ten / boar / to / (Shang)Chia 114 (10) (It is) ten boars that we should (yu-offer) to (Shang)Chia. - 231 -In this group of inscriptions, there are in fact two OBs, i.e., (Shang)Chia ( J> ) xf and the Pao-ances tors (Pao I ^/sJ. 2- , Pao Ping ^ ^} and Pao Ting "[ ). However, the R-S activities related to these groups of ancestors do not appear to be in coordinative relation ship. The clause 1 ^^-^^ ' is rePeated without any change in both inscriptions, while the number and type of victims to be sacrificed to Shang Chia change. This shows that '-j^^ ^1 is something already determined; the real question is what should be yu-offered to Shang Chia. (The deletion of the clause ' % fifi ' in 114 (10) also suggests that it is 'old information'.) Such a practice may be similar to the p'ei ssu t>\\j ('paired sacrifices', a notion introduced by Shima 1958: 201). 'Paired sacrifices' seem to be a practice of long tradition in Chinese rituals. In Li Chi ^Jf ff^j , it is stated that: (SSCCS, chuan 24, p.2) ... when the people of Lu are going to sacrifice to the ti-god, they will, for sure, sacrifice in advance, at the Pan Kung; when the people of Chin are going to sacrifice to the River(god), they will, for sure, sacrifice in advance at 0 Ch'ih (or Hu Ch'ih); when the people of Ch'i are going to sacrifice to the (spirit of the) T'ai Shan, they will, for sure, sacrifice in advance at the P'ei Lin.' How this p'ei ssujffit, %\j was practiced is something we know very little about. How the 'partner' ffyt> sacrifice was determined and the nature of the relationship between this partner and the major receiver of a sacrifice are questions we are unable to answer. In the case of the ritual kao Jj^£ , there are some other inscriptions which also - 232 -suggest a p'ei ssu-like practice: keng-yin / test / perhaps / announce / high £ 11 * * f * ®* Ancestor / burn-sacrifice / to / (Shang)Chia 5 = f * three / ox Nan Ming $fa ^j) 470 If we perhaps make an announcement to the great ancestor(s), we should burn-sacrifice to (Shang) Chia three oxen. Or: We should perhaps announce to the great ancestor(s) and burn-sacrifice to (Shang) Chia three oxen. hsin-ssu / crack / perhaps / announce / water / enter 0t * t& ^ ^ is to / Shang Chia / pray / T'ai I / one / ox / king receive / assistance Ts'ui 148 If we announce the flood (lit. that the water entered ...) to Shang Chia and pray to T'ai I (with) two oxen, the king will receive assistance. kuei-ch'ou / crack / Pin / test / ju-sacrifice / T'ai Chia announce / to / Ancestor I / one / ox / eight / month / use I 4^ 115 We should yu-sacrifice to T'ai Chia and make an announcement to Ancestor I (with) one ox. One the eight month, they are applied. / - 233 -In the inscriptions Ping 114 (8) (.9) , the OG (the matter being announced, the content of the announcement) is not made explicit. If the graph (word) really means 'shackled captives (in a military expedition)', the activity kao Jt^ may be a ritual performed to announce a Shang military victory. The performing of the kao ritual is accompanied by the presenting of captives. In Chapter I, we have already cited the passage from the Chu Shu Chi Nien Vj-^" |£ 1^ : ' V,A )^f1 (he announced (his victory) by presenting the captives and the left ears of the slain' (Chapter I, footnote no.4). We find a similar description in the Li Chi; %fo , %k ^ fa. , ff j| J ff, ^ H ftf (SSCCS, chuan 12, p.4) They set out on a punitive expedition, capture the wrong doers and return. Then they present vegetables and cloth in the school and announce (their victory) by presenting the captives and the left ears of the slain. On the same plastron with Ping 114 (8), (9) and (10) we find t9 f t i H ty$ nrf* %*i chia-ch'en / crack / Ch'eng / test / we /attack I3A n f*£\*i \* Ma Statelet / ti-god / give / us / assistance 114 (7) If we attack the Ma Statelet, ti-god will give us assistance. This inscription seems to argue for the contextual interpretation proposed above. (.3) Ping 21624 test / ti-sacrifice / to / east / direction / call / Hsi wind / call / Hsieh / invoke / harvest 216 (3) - 234 -We should perform a ti-sacrifice to the (spirit of the) east quarter called Hsi and the (spirit of the) wind called Hsieh to invoke for a (good) harvest. 4. test / ti-sacrifice / to / west / direction / call / I wind / call / Han f invoke / harvest 216 (4) We should perform a ti-sacrifice to the (spirit of the) west quarter called I and to the (spirit of the) wind called Han to invoke for a (good) harvest. hsin-hai / crack / Nei / test / ti-sacrifice / to / south direction / call / Wei / wind / Jen(?) / invoke / harvest 216 (2) We should perform a ti-sacrifice to the (spirit of the) south called Wei and to the (spirit of the) wind (called) Jen(?) to invoke for a good harvest. Ilh :^ Wi ?4 fj K*-hsin-hai / crack / Nei / test ? ti-sacrifice / to / north f % i^v? /7\. ufc) A fH eK? \ ^ £ direction / call / Yu"an(?) / wind / call / (?) / invoke harvest 216 (1) We should perform a ti-sacrifice to the (spirit of the) north quarter called Ytlan(?) and the (spirit of the) wind called $ C?) to invoke for a (good) harvest. - 235 -Shima has made a detailed study of the ti-sacrif ice .'0958:198-218) . His study has shown that tl^ is a sacrifice related directly or indirectly to agriculture, e.g. an invocation for rain or a ritual performed to the wind gods. The author has reservations about accepting Shima's theory that the ti-sacrifice is performed to the ti-god since many spirits, e.g.-^" Mountain (god) , "^/rfj River (god), earth (god), Wang Hai Jx (an ancestor), Shang Chia ^ , Hsia I , appear as OBs of this sacrifice. Although Shima applies the notion of p'ei ssu fyftL%lLito explain this phenomenon, he has not demonstrated that in inscriptions such as ' ""jjjj ^ -^h ' (I 846), the ti-sacrif ice is performed to the ti-god who is the main OB while is only the minor OB (the partner of the ti-god). (For a review of the various scholars' theories concerning the ti-sacrif ice, see Ikeda 1964:89-93. The author agrees with Ikeda in rejecting Shima's theory of p'ei ssu.) On the other hand, it is also very difficult to trace the assumed evolution of the ti-sacrifice from the O.B.I, to the classics. As noted by Shima, there is a great deal of controversy about the character of theVy (.1958:212). Without sufficient knowledge of the >?5j5 f\ in the classics, we can hardly draw any comparison between it and the hjp ti-sacrifice in the O.B.I., although similarities of graphic structure (and, presumably, pronunciation) suggest an evolutionary development from one ' ' ' V to the other. - 236 -Confining our study of the <"tj? ti-sacrifice to the O.B.I-.,, we find the verb t_i shows a dual nature. There are some inscriptions where _ti /-^j shares the characteristics of type B verbs, i.e., it signifies a specific method of sacrificing victims: test / fang-sacrifice / ti-sacrifice / one / Ch'iang ^> - -two / dog / dismember-sacrifice / one / ox Ping 122 (9) We should fang-sacrifice and ti-sacrifice one Ch'iang-tribesman, and dismember one ox. (Notice that there is still an OV 'Ch'iang ^ " following the verb _ti although there is a type B verb mao ty^ .) wu / crack / fang-sacrifice / ti-sacrifice / three / pig •±••4 t* fj JL*. yu-offer / dog / dismember-sacrifice / to / Earth(god) penned-sheep / invoke / rain I TA 40 We should fang-sacrifice -and ti-sacrifice three pigs, yu-offer one dog; and dismember-sacrifice to the Earth(god) one penned-sheep to invoke for rain. (It is not clear whether the Earth(god) ^— is also the OB of the sacrifices fang ~$ and ti_ 1p or not. Syntactically, it is possible to take T'u — as the beneficiary of all the sacrifices, i.e. fang /) , ti ^> , yu and mao . Nevertheless, a sentence such as '* -jf X,. to ti-sacrifice to the Earth (god) ' never appears. Shima contends that T'u is not one of the OBs to whom a ti-sacrifice is performed (1958: 202).) - 237 -On the other hand, ti ^ sometimes exhibits linguistic behaviour similar to that of type A verbs: test / ti-sacrifice / to / east / bury a dog k'uei-kill-sacrifice / pig / burn-sacrifice / three penned-sheep / dismember-sacrifice / brown / ox Hstiff 2.18.B In performing a ti-sacrifice to the (spirit of the) east, we should bury a dog, k'uei-kill-sacrif ice a pig, burn-sacrifice three penned-sheep and dismember-sacrifice a brown ox. (Notice that the word tjL W does not directly take an OV.) test / fang-sacrifice / ti-sacrifice / dismember-sacrifice one / ox / yu-offer / one / piglet Ch'ien 7.1.1 In performing a fang-sacrifice and a ti-sacrifice, we should dismember an ox and offer a piglet. (This inscription is similar to the one above in that ti_'«p does not directly take an OV. However, as this inscription appears on a small bone or plastron piece (see the attached copy), some graphs may be missing Therefore, this inscription cannot be adduced as solid evidence.) Ch'ien #j 7.1.1 - 238 -Unfortunately, both types of inscriptions pointing to lip as type A and type B occur in small number. We cannot ascertain whether such a peculiar linguistic behaviour is the idiosyncrasy of the word ti or whether perhaps scribal errors are involved. For the time being, we tentatively classify J^L rQj as a type B verb representing an R-S activity accompanying the ritual ch' iu ?ft . (.4) Ping 166 (5) f t rf ** A* call upon / yu-sacrifice / present Interpretation A: We should call upon (so and so) to perform a yu-sacrifice and a present-ritual. Interpretation B: We should call upon (so and so) to yu-sacrifice (in order to) perform a present-ritual. Interpretation A which treats %ffy and as two coordinate verbs is the most simple, direct analysis and is syntactically acceptable. Nevertheless, interpretation B is proposed as an alternative which seems to reflect the underlying relationship between these two sacrifices more precisely. In the above sections, we have already classified ^(5) as a type B verb. Now, let us look at the verb §L . There are many variant forms of the graph, (word) teng g , but we may consider ^ \ (two hands holding something to be presented) and .Su (a food vessel) as basic elements. While there are different - 239 -transcriptions proposed for this graph(word), most scholars agree in interpreting it as 'to present' (Li Hsiao-ting; 1965:1673-77). Such an interpretation is well supported by its graphic structure, its usage in O.B.I, contexts and its later equivalent in the classics. Such a meaning naturally leads us to interpret the words such as ^ 'millet' and ^7 'aromatic wine' which frequently follow as direct objects(patients). Accordingly, we may put it in the same category with words liao ^ , yu and yu ^ , since the parallelism between ' ^ / ^f&) / ^ + OP' and ' ^ + ' is very evident. It appears that teng is also a type B verb. However, there are some peculiarities of this word: (1) Unlike the words liao , yu fyfy and yu whose direct objects usually appear after the OB, i.e. ' / I + ( ) + OB + OP', the direct object of teng invaribly appears in front of the OB, e.g., rJS 'hSi T* f% test / perhaps / present / grain/ to _ j 7-1 L. Ancestor I Nan Nan ffiffil. 136 ... we should perhaps present grain to Ancestor I. kuei-hai / crack / Ho / test / perhaps / present aromatic wine / to / Ancestor I / hui / next W ft a. i-ch'ou Chia ^ 2407 If we perhaps present aromatic wine to Ancestor I, (this ritual should be performed on) the next', i-ch'ou day. - 240 -kuei-wei / crack / present / wheat I to I two tablet K/u ft 1061 We should present wheat to (the spirits of) the two tablets (Shih Jen tf\ ^ and Shih Kuei & A ?). (See also Ts'ui-j^ 269, Ching %. 4025, Ts'ui ^ 908. List inexhaustive. In the Sorui, there is no case of '*^+ (?) + 0B+ $ )'•) The pattern ' + ^ / 7f\ / ^ / '% + ( -f ) + OB* is parallel to the pattern + OG + ( ^ ) + OB. ' Therefore, the objects (^ / ^ / / ^ ) appear to be more like OGs than OIs. (2) In the inscription To ^4 1.457, the graph(word) teng exemplifies linguistic behaviour parallel to that of kao , yu and ch'iu >|c . i-hai / crack / present / aromatic wine / three Ancestor Ting / penned-ox / king / receive / assistance If we present aromatic wine to the three Ancestor-Tings (with) a penned-ox, the king will receive assistance. The structure ^ + % + (. tf ) + £ + ^ 1 is identical to that of ' ^ +$f + + '>it' and ' j^L + ^ + f + + ^ ' (all cited on p.200.) - 241 -If both ^jjf 'aromatic wine' and ^ 'penned-ox' function as the patients of the verb ^ 'present', why are they separated by the beneficiary J 'three Ancestor-Tings'? Why doesn't it take the pattern '* % C|/ f J £ & j"' or •* f 5. *E J $ ^ ' ? In looking for an explanation, we turn to the Li Chi (SSCCS, Li Chi, chUan 16,p.5). " ... then the peasants present millet. In this month, the emporor tastes the millet with a chicken; and offers it with cherries, presenting it at the ancestral temples. ^ K % fa , £ % rt ft (ibid. p.23) to taste the hemp(?) with a dog; and present it at the ancestral temples. ( ib id. chUan 17, p. 7) ... then the emporor tastes the rice with a dog; and present it at the ancestral temples. The commentary on the sentence 'i*/^ ' is very illuminating; it states, 'it means to taste the piglet. The reason why it says yi  ch'ang shu is that the main concern does not lie on the chicken, instead, it lies on the grain.' 0*fc$Bt*A<f£ ,7*'Ml ,i %L4L> (ibid, chdan Y6, p.. 5) This emphasis on the grain is easy to understand. In an agricultural society where even the emperor has to do some symbolic farming, grain products enjoy an unrivalled importance. The following passages from the classics can illustrate this: - 242 -(Kuo YU Yin Te (jj) j/ 16b//18) The emperor hoes one section of earth; the officials, according to their ranks, triple the hoeing; and the masses finish (hoeing), one thousand acres. (Mencius HY 22/3b/3) It is said in the Book of Rites, "A prince ploughs himself, and is assisted by the people, to supply the millet for sacrifice. /T 0^-,s (Legge, p.266-267) ^ tt. (SSCCS, Li Chi, chflan 17, p.2) 'present the best of the five grains, and store the product of the king's farming in the sacred granary in a respectful and solemn manner.' Cheng HsUan glosses, (ibid.) '(it means) they are serious about the storage of the millet for sacrifice ... that which is used for the storage of the sacrificial millet is a sacred granary.' All these passages refer to the practices of the Chou Dynasty, but there seems to be some evidence suggesting that the Shang people had similar rituals. king / should not / noe ... Ping 479 (.3) The king should not hoe ... - 243 -Sil* 1 c ii &** IfM chi-hai / crack / test / king / go / observe hoe Chia ^ 3420 The king should go to observe the hoeing. chi-ssu / test / king / perhaps / present / south ©act) ^* f * granary / grain / hui / i-hai If the king is going to present the grain of the south granary (this sacrifice should be performed) on the i-hai day. Chia ^ 903 chi-ssu / test / king / grain / store / perhaps ft*, ft, #*fi li> present / to / Ancestor I The king's grain is stored, we should perhaps present it to Ancestor I. Chia ^ 903 Of course, it is premature to contend that the millet-sacrifice recorded in the Chou classics was practiced in exactly the same way in Shang times. Nevertheless, the numerous inscriptions concerning rain and the soliciting of good harvests suggest that the Shang was an agricultural society. By drawing comparisons from the classics, we may hypothesize that grain and related products (e.g. ^ ) were sacrificed not so much as a means of soliciting the ancestor' s assistance when needed, but rather as a moral obligation, a ritual performed for its own sake. In other words, to present grain or beverages - 244 -made with grain is itself a goal. In this sense, the words ' ^ 1^1 appear as OGs similar to the word J^" 'harvest' in the structure jfy ' to invoke for a (good) harvest'. Hence, these grain products do not function as a means (01) of soliciting a specific favor from the ancestors. Instead, they take separate 01s, e.g., ^ ox» to accompany the presenting of the grain products. The sentence '^v^ J"^ , should be interpreted as 'if, in presenting the aromatic wine to the three Ancestor-Tings, we USE a penned-ox (to go with the aromatic wine), the king will receive assistance.' The following two inscriptions can also be interpreted in the same way: perhaps / present / new / aromatic wine / hui two / ox / use Ts'ui Tf 910 If we perhaps present the new aromatic wine, it is two oxen that we should use (to accompany the sacrifice). (Cf. on the same bone/plastron piece, there is an inscription ' ^ "*f ^ tfj it is a penned-ox that we should use'.) present / millet / ... / sui-cut-sacrifice penned-ox Ts'un ff 1.1769 ... in presenting the millet, .... should sui-cut-sacrifice a penned-ox. (This inscription seems to reflect the same sort of sacrificial procedure as found in the sentence %, ' from the Li Chi.) - 245 -After illustrating that teng jgf behaves like a type A verb whose performance requires another sacrifice to accompany it, we may again look at the inscription ' ^ ^ ' (p.238). Is it possible to interpret it as 'we should call upon (so and so) to yu-sacrifice (in order to) perform a present-ritual'? Compare the following inscriptions: 1* ft*. H§ i& U As f wu-ch'en / test / invoke / grain / from / (Shang)Chia perhaps*/ burn-sacrifice Jin A 2363 In invoking for a (good) harvest (starting) from (Shang)Chia, we should perhaps perform a burn-sacrifice. ting-wei / crack / perhaps / invoke / harvest / to River(god) / hui / hsin-hai / yu-sacrifice If we perhaps invoke for a (good) harvest to the River(god), it should'be on a hsin-hai day that we perform a yu-sacrifice. Chia ^ 1885 ¥? 1* 1* I ** hsin-wei / test / invoke / grain / high / ancestor ii «i 7> ?| ^ River(god) / at / hsin-ssu / yu-sacrifice burn-sacrifice Chih HsU ^ ^ 2 If we invoke for a (good) harvest to our grand ancestor — the River(god) (or to our grand ancestors and the River(god)), we should, on a hsiri-ssu day, perform a yu-sacrif ice and a burn-sacrifice. - 246 -present / aromatic wine / Father Chi / hui this / day / chi-hai / yu-sacrifice Jin /\ 1810 The parallelism between these four inscriptions is so obvious that we can with confidence interpret the last inscription as '... in presenting aromatic wine to Father Chi, it should be. on this chi- hai day that we perform a yu-sacrifice.' That is, the type B R-S activity Is performed, to accompany or to realize the type A R-S activity v£. . When the type B verb ^] is placed in front of type A verb ^ , i.e., ^] ^ , we can justificably interpret it as 'to yu-sacrifice (in order to) perform a present-ritual'. Accordingly, we may interpret the following inscriptions in the same way: ( r& 1 K jf fw h ch'ou / crack / Pin / test / next / i-... / yu-sacrifice Ti kit I h millet / present / to / Ancestor I / / king / prognosticate ^ $4 ;H UJL rn^n^R^ say / have / harm / not / perhaps / rain / six / day ... •wu / night / moon / have / eclipse / i-wei / yu-sacrifice 4 M many / kung-official / ? -*/• ->.l ? * Ping 57 (1) On the next i_-... day, we should yu-sacrifice to ^present.".millet1, to Ancestor I. The king prognosticated and said, 'there will be harm, it will perhaps not rain. After six days, on a ... wu night, the moon was eclipsed. On the i-wei day, we performed the yu-sacrifice. Many kung-officials (artisans) ^p$>i^ (?)•' 26 - 247 -I chi- / ... / crack / .. . / test / should not fi if , A t yu-sacrifIce / present Ping 57 (2) We should not yu-sacrifice to perform a present-ritual. (For the interpretation of Jt kung, see Hsiao Nan 1981.) - 248 -III. THE SIMULTANEOUS-SUCCESSIVE SENTENCE We have already investigated two types of structure for which we proposed the analyses: 1. 'In doing such and such, we should do thus and so' 2. 'We should do thus and so(in order to do such and such' Although difficult to identify in many cases, there is the third structure which can be analysed as 'after we do thus and so, we should also (or, we should proceed to) do such and such.' For example, test / king / all / yu-sacrifice / present / should not treat as guest / i-.j ih ritual Ping 34 (3) The king has all completed the yu-sacrifice and the present-ritual, he should not treat (Shang Chia) as a guest during the i-jih ritual. tff fl* 1 i*_ i h P> A? M$ jit , „ 4 chia-ch'en / crack / Ch'Ueh / test / king / treat as guest n a B i-j ih ritual Ping 34 (4) The king (has completed the yu-sacrifice and the present-ritual), he should treat (Shang Chia) as a guest during Vthe i-jih ritual. (The restoration of the OB (Shang Chia) in the translations is based on Ping 36 (3), (4) which is a member of the same set - 249 -C -$\ ) as Ping 34 (3) and (.4). In Ping 36 (3), the name 'Shang Chia' appears.) In the O.B.I., the pin ritual co-occurs with various sacrifices. Although we do not know how this ritual was practiced nor exactly what procedures it incorporated, there is one passage in the Shang Shu which provides some hints: yk £ #r ft at % k t * A* * - ^u 27 (SSTC 330734) On the day Mow-shin, the king in the new city performed the annual winter sacrifice, offered a red bull to King Wan, and the same to King Woo. He then commanded a declaration to be prepared, which was done by Yih in the form of a prayer, and it simply announced the remaining behind of the duke of Chow. The king's guests, on occasion of the killing the victims and offering the sacrifice, all made their appearance. The king entered the grand apartment, and poured out the libation. (Legge, p.451-2) Legge's translation is based on the traditional commentaries. Our understanding of the O.B.I, language leads us to propose some revisions to his translation.. CD Sui Legge notes, 'Ts'ae read the three characters together, with the meaning which I have given in the translation;,— whether correctly or not I cannot undertake to say. The , occurring where it does, is a great difficulty. Lin Che-k'e says the best plan is to allow that the sV,' is - 250 -inexplicable, and so pass over it. I believe he is right. (ibid. 452) Legge is justified in casting doubt on the traditional interpretation. In fact, the word sui appears to be a specific method of killing and sacrificing victims, e.g., >f s. ft *t |+ Ii ^ &>t'J) yu-offer / behead / ten / five / sui-cut-sacrifice small / penned-sheep / (Shang)Chia I_ IA 78 We should yu-offer beheaded-victims amounting to fifteen and sui-cut-sacrifice a small penned-sheep to ($hang) Chia. ping .. / crack / test / Pi / perform libation rf * °U Ir* sui-cut-sacrifice /'Ch'iang / thirty / dismember-sacrifice three / penned-sheep / split-sacrifice / one / ox / at f % Wit) /N A j f\ ancestral temple / use / eight / month Hayashi ~Jf%\ 2.3.11 Pi should perform a libation, sui-cut-sacrifice Ch'iang-tribesman amounting to thirty, dismember-sacrifice three penned-sheep and split-sacrifice one ox. (These R-S activities should be) practiced in the ancestral temple. In the eighth month. The parallelism between the verbs sui , yu , mao &f and fu obviously suggests that sui is also a specific method of killing sacrificial victims. T'ang Lan interprets it as a phonetic-loan for sui jfc'|''to sui-cut' (1939:28). The Shang Shu sentence '^'$~ ^ — --' - 251 -is comparable to the following O.B. inscription: n(% Fr* u ff $ ft ping-ch'en / crack / sui-cut-sacrifice / to Ancestor Chi /ox Ho ^ 149 We should sui-cut-sacrifice to Ancestor Chi one ox. (2) £in ^ Legge translates the structure 1 -«S 'as 'the king's guests', a subject taking ' ifyl all made their appearance' as a predicate. However, numerous O.B. inscriptions suggest that ' ^ ^ ' is a 'subject + predicate' structure, i.e., 'the king treats ( — ) King Wen and King Wu as guests' (Ch'en Chih , 1930:3) (see also footnote no. 29). And accordingly, the expression ' "f^ ' may be interpreted as 'the ancestors all came (as guests ).' cf. (SSTC 05^gQ) 'the imperial progenitors come to the service'(Legge, p.87) . The ritual of treating spirits as guests is also encountered in the Yao Tien^-Sfe Chapter, ' £ ^ £ t3 ' (SSTC 01Q009) respectfully to receive as a guest the rising sun' (Legge, p.19). In the Li Chi, there is another indirect reflection of this practice ' 4C"(^ K^T? ^ jf^: X_ > (SSCCS, chuan 42, p.11) '(in a funeral), you treat your (deceased parents) as guests, that is why it is sorrowful'. If we apply the interpretation appropriate to the O.B.I, proposed in footnote 27 to this sentence from the Shang Shu, we may translate ' ^ as 'the king treated (King Wen and King Wu) as guests in perform ing the killings and sacrifices.' - 252 -The Shang Shu passage suggests that there are R-S activities performed before (i.e. cheng ^ , chi ^ , sui JpjJ )and after (i.e. kuan ^-f; ) the pin /f^ ritual. Keeping these practices in mind, it appears that we may analyse the sentences ' 5 ^ ^/ftf ^ > ^ j| ^ 0 ' and 1 £ /§ ^ fl ' as reflected in the proposed translations (see p.248) Also, according to the I Li /f^_ , some rituals take days to be completed. The ritual t'e sheng k'uei shih ^ ^ ^|[ ^ is an example. flM^B iLm,%r (SSCCS, chUan 44, P.4) IffL *8 ? (ibid. p.7) /*l ^ (ibid. p.9) 'Three days before the scheduled sacrifice, use milfoil to divine about who should be the shih (a person who represents the deceased ancestor).' 'In the evening of the next day (after receiving the guests participating in the sacrifice)' 'Get up in the morning ...' Although we cannot project back unconditionally these procedures recorded in the I Li to the Shang period, it seems reasonable to assume that the Shang, who were known for their multifarious R-S practices, also spent days in religious activities. The numerous victims used in one sacrifice already suggests that their R-S rituals are very time-consuming and labour-intensive. (See Chang Ping-ch'Uan's £^ 'Chi Ssu Pu Tz'u Chung Ti Hsi Sheng' fjjfc fa \. j|| ff7 1968.) - 253 -Thus it appears occasionally to have been the case that, during the proceedings of an R-S activity, the Shang divined about whether they should go on to the next step or not. X* 9$ ^ $, fl* ¥a kuei-yu / crack / /£-(?) / test / next chia-hs{f~ / finish / yu-sacrif ice / hsieh-ritual from / (Shang)Chia / i-sacrifice / ... / to many / king T'ung (jfg_ XI 2 On the next chia-hsU day (the next day of kuei-yu), after finishing the yu-sacrifice, we should perform a hsieh-ritual starting from (Shang)Chia, (and then) 28 perform it to the many (deceased) kings. kuei-hai I ... I chia-tzu / finish/ yu-sacrifice i-j ih / from / Shang Chia / ^-sacrifice / arrive to / many / king / have no / misfortune / three / month If, on the chia-tzu day (the next day of kuei-hai), after finishing the yu-sacrifice, we perform an i-j ih ritual starting from Shang Chia (and then) perform it to the many (deceased) kings, we will have no misfortune. At the third month. Ts 'ui85 (See also Chin ^ 124, Ming f>fj 307, K%a/j| 1230, Hayashi ^ 1.21.7 and Hstt Ts'un ^ ^" 1483. List inexhaustive.) - 254 -It seems justified to interpret the inscription ' ^ ^ > v '/l ' 13 1 as a successive sentence. Concerning the interpretation of ' j£ ^ 0 (or ^ £?0 1? )', some remarks must be made: In the O.B.I., the word pin frequently occurs in the pattern ' ^ + A + ^ + s'" ^here are two possible ways of analysing this pattern: (1) Pin ^ is a major ritual (type A) the performance of which requires the accompaniment of other R-S activities, i.e., 'in treating so and so (the spirit of an ancestor) as a guest, we should perform this or that R-S activity.' (2) Pin % is an optional ritual/procedure (type B) in the A performance of other R-S activities, i.e., 'the king should treat so and so (the spirit of an ancestor) as a guest 29 during the performance of this or that R-S activity.' Although differing somewhat from the first analysis proposed above, Ch'en Meng-chia's interpretation treats pin ^\ as a major ritual (1956:100). He analyses the above pattern as 'S - V - 0^ - 0^' (S stands for the subject) V, verb; 0^, indirect object; 0^, direct object.) Unfortunately, he fails to elaborate on how the verb pin ^ takes an R-S activity as a direct object and the spirit of an ancestor as an indirect object. Presumably, he treats the structure ' ^ + ancestor + Vp c' in the same way as he treats the structure' ^ < for 'preposition'). Notice that Ch'en mentions that in the pattern ' (Ch'en's analysis V - P - 0^ - 0^ where P stands - 255 -' + ancestor + VR g', the preposition yjl ij- before the indirect object (ancestor) is omitted (ibid.). That is, according to Ch'en, the pattern ' + ancestor + V ', in its hypothetical full form, yi R—s would show the structure '* 'j*^ + -f + ancestor + VR_g (V - £- 0^ - 0^) which is identical to that of ' %j~> ^ £ 7V*^ ' (to invoke to Wang Hai (with) nine oxen). As a matter of fact, it is very natural to interpret pin SK) as a major R-S activity. Pin /ft\ is one of the most common R-S verbs in the O.B.I., it co-occurs with more than forty R-S activities in a pattern identical or similar to the one cited above (Shima counts fifty R-S verbs which co-occur with pin (1958:314-315). An R-S activity of such a frequent occurrence naturally impresses us as very important and we might easily be led to interpret it as a major R-S activity requiring the accompaniment of other R-S activities. However, such an interpretation runs into difficulties when we encounter the following inscriptions: ' keng-tzu / crack / test / Ancestral Mother Keng cf fc It US, %\ sui-cut-sacrifice / king / perhaps / treat as guest <^ In performing a sui-cut-sacrifice to Ancestral Mother Keng, the king should perhaps treat (her) as a guest. (h ^ Ming Pfl 69 should not / treat as guest Ming &f\ 69 The king should not treat (her) as a guest. (See also Ch'en 34, Jin K 1550, Ts'un fy- 2.600.) - 256 -test / Hsiao Ting / sui-cut-sacrifice / perhaps treat as guest I l\ 850 In performing a sui-cut-sacrifice to Hsiao Ting, we should perhaps treat (him) as a guest. ^ test / should not / treat as guest We should not treat (him) as a guest. I 850 Obviously, the R-S activity sui is something the Shang have already decided to do, so the question is whether, during the sui-sacrifice, they should, as an option, treat the ancestor in question as a guest. These inscriptions seem to refute the theory that pin >n is a major ritual. One may argue that these inscriptions can be interpreted as successive, rather than conditional sentences, i.e., 'we have already performed the sui-cut-sacrifice (as a preparatory procedure), we should/should not proceed to perform the (major ritual) pin.' Admittedly, such a possibility exists as far as the above two pairs are concerned. However, such an interpretation does not work in the following case: wu-hsd / crack / perhaps / shih-sacrifice (?) 5 7. f * f. t. to / Ancestral Mother Chi / king / treat as guest If we perhaps perform a shih-sacrifice(?) to Ancestral Mother Chi, the king should treat (her) as a guest should not / treat as guest (The king) should not treat her as a guest. Ning & 1.218 Ning % 1.218 - 257 -The occurrence of the modal ch'i shows that, at the time when the divination is made, the shih-sacrifice has not been performed. Therefore, the successive interpretation must be excluded. Consequently if we adopt the conditional interpretation as reflected in the translations, it seems pin can hardly be interpreted as a major R-S activity. In inscriptions Ping 34 (3) , T'ung (JQ^ XI 2 and Ts 'ui 85, we find the words hsien Jj^, and ch' i \ signifying the completion of an R-S activity. Nevertheless, these words, not to mention aspect markers such as yi ^ , are not always employed. There are also cases where, by context, we can tell that some event has already taken place, even though the inscription does not include a temperal marker. kuei-mao / crack / Cheng / test / king / order three / earl / shoot / not / announce / ten T tt • S © : .it tablet / king / misfortune / foei / this Ping 83 (1) The king ordered three earls to shoot (or, ordered three hundred shooters), he did not announce this activity to the (spirits) of the ten tablets. His misfortune is (caused by) this. (.I.e., the king is obliged to announce/report his activities to the spirits, but as he failed to do so, the spirits have punished him.) - 258 -test / king / misfortune / not / wei / this not / announce / three / earl / shoot The king's misfortune is not because < of this; he failed to announce that (he ordered) three earls to shoot (or ordered three hundred shooters) 83 (2) On the same plastron, are found the inscriptions: test / king / misfortune / favorable 83 (7) The king is experiencing misfortune, (he)'will be all right. ^Ai tin y# £ fie ft) king / misfortune / not / perhaps / favorable The king is experiencing misfortune, (he) will not \b& all right. ( 83 (8) The last two inscriptions show that the king is experiencing misfortune. And the first two inscriptions are divinations made to find out whether or not the king's misfortune is due to his failure to report his activity (4^5- "E( *f& ) Hf ) to the spirits. Notice that the interpretation of ' ^ > J§ ( f£) ), "n. "X 3f~* ' as activities already taken place is based on considerations of context. (The interpretation '*if the king orders three earls to shoot (or orders three hundred shooters) and if he does not announce it to the (spirits of) the ten tablets, . his misfortune will be due to this (failure of announcing)' sounds very strange. Cf. According to Ch'en Meng-chia's - 259 -study, the word chih (this) in the expression chih jih ^ Q 'that day', chi hsi ^ 'that evening' refers to a date in the past (1956:114) since / often occurs inj^J?j(/fjjji describing what actually happened. '•• Notice the *p- type negative fu .) Also in the verifying statements (yen-tz 'u.E^A *<>] )which refer exclusively to activities or events in-the past, we do not always encounter words (or markers) signifying completion. The paucity of these words prevents us from offering definite interpretations for many inscriptions. Therefore, the inscription ' (cited on p. 238) can also be interpreted as 'We have already called upon (so and so) to yu-sacrifice, (now we should proceed to) perform the present-ritual.' This is exactly the same problem we encountered, in trying to distinguish a simultaneous-successive clause from a conditional clause (see Chapter III, p. 142). In the cases where there, is no negative counterpart something like '* *£, ' (*we should not call upon (so and so) to yu-sacrifice to accompany the present-ritual), we cannot exclude the poss ibility that is a simultaneous-successive sentence. For these reasons, the following pair of inscriptions is, syn tactically, open to two interpretations: king / perform libation / ting-ritual yu-offer / behead Ping 122 (5) - 260 -, The king has already performed a libation, he should also (or he should proceed to) perform the ting-ritual (a ritual performed by making use of a cauldron (?)) and yu-offer a beheaded victim. Or: The king, when performing a libation, should also perform a ting-ritual and yu-offer a beheaded victim. king / perform libation / should not / yu-offer it ti behead Ping 122 (6) The king has already performed a libation, he should not also yu-offer a beheaded victim. Or: The king, when performing a libation, should not also yu-offer a beheaded victim. (For the interpretation of the verbal usage of ting ^ff , see Takashima 1981.) As in many other cases (see footnote no.20), it is difficult to explain why the word ting is omitted from the negative counter part. Presumably, after making the first divination, the Shang have obtained some 'instructions' from the turtle plastron and have decided whether or not they should perform the ting-ritual. Therefore, there is no longer any need to divine about its performance. Or, as Takashima proposed, ting functions as an adverb decidedly used to counter-balance the strong force of wu ^ (ibid. 21). / - 261 -As mentioned in footnote no.l, the word kuan never takes an OV. Therefore, the R-S activity ' ^ ' cannot be analysed as a specific method of sacrificing victims to realize the ritual kuan ^JfL Also, it has been shown that y_u ^ and kuan "j-^ can appear in coordinative or contrastive relationship (see p. 222 ). Thus, kuan and y_u may be interpreted as parallel rituals which are parts of a certain R-S activity unexpressed in the inscriptions. If we adopt the. first interpretations, we may hypothesize that these divinations were made after the first procedure kuan had been performed and the Shang had divined about whether or not they should proceed to do the following rituals. The descriptions in the Li Chi imply that the kuan Jfrffi ritual precedes the victim sacrifice in certain observances. (SSCCS, chuan 49, p.5) The king holds a spoon made of the kuei-jade to perform the libation to the shih•(the representative of the deceased ancestor); the ta-tsung (the king's wife) holds a spoon made of the chang-jade to assist the libation and (then) receives the victim. Also the passage (ibid., chuan 21, p.16) 'to use the hsUan-wine (water) to perform the ritual, (then) present the blood and the hair (of the victims)'. See also the commentaries (ibid., chUan 21, p.14-15) where a detail description of the ritual procedures are given. Or, if we adopt the second interpretations suggested above, we may hypothesize that the Shang have already decided they are going - 262 -to perform a libation and the question is what other R-S activities should also be performed. Cf. the Li Chi states, fok i/fb iUffi' Cibid., chUan 49, p.7) 'there are thrjee steps in th_e rituals (presenting > music /*gf, and dance 'JJK ), among the presentings, there is nothing as important as kuan.' Although we cannot insist on the application of these Li Chi descriptions to the O.B.I., they seem useful as references. Unfortunately, due to the lack of formal markers, we cannot determine which interpretation is correct and must treat them both as possibilities for the time being. - 263 -CONCLUSION In the previous chapters, we have investigated three types of subordinative composite sentence, namely, 'cause and effect', 'conditional' and 'simultaneous-successive'. As observed, there are no formal connective markers, so the logical relationship between two clauses can only be determined on the basis of semantic considerations, the tui-chen pair, the practice of abbreviation and the larger context. A major type of 'cause and effect' sentence is the sentences of the pattern ' ^ + v rf I ^ ...' where the second clause represents some undesirable effect or situation. A consideration of the general positive versus negative pattern of the O.B.I, and the practice of abbreviations has led us to adopt the analysis 'cause and effect' for this sentence-type. Due to the lack of connective markers, we cannot distinguish a conditional clause from a simultaneous-successive clause in many cases. Therefore, these two types of sentences are treated together in one chapter. We may interpret sentences in which ch'i ^ appears as conditional. Nevertheless, rather than a pure subordinate marker, the word ch'i is interpreted as a modal conveying the sense of uncertainty, a usage well illustrated in the classics. - 264 -Also, the theory that treats ch'i as a marker of an embedded sentence has been refuted because it does not seem to fit the general syntax of Chinese very well and it requires the unjustificable assumption that the focus of a ming-tz'u /»p ^ is shifted in a prognostication. The apodoses of 'conditional' and 'simultaneous-successive' sentences may represent an intended result or an undesirable consequence/situation. In most cases, these two types of apodoses can be easily distinguished. But in the case of raining, we have to rely on the idiomatic expressions ' % {fa ( t li? )' and ' $t_(fa ( )' in drawing the distinction. In divining about the appropriateness of a proposed activity, the 'conditional' and 'cause and effect' sentences both serve the same purpose. But an activity of greater gravity seems to motivate the employment of the latter. The R-S verbs can be roughly divided into two categories, i.e., type A and type B. Type A verbs represent major R-S activities requiring the accompaniment of type B R-S activities. The type B verbs can be placed either in front of or after the type A verbs. The latter case constitutes a conditional or simultaneous-successive sentence, while the former constitutes either a composite sentence incorporating a 'to clause' or a simultaneous-successive sentence. - 265 -NOTES FOR CHAPTER ONE The graph is composed of two elements: £j and , which is graphically equivalent to the graph . But to interpret it as 'g^ does not make any sense in the context. Lu Shih-hsien i^, proposes that *grwan be understood as a phonetic-loan for ^ *kwan and interprets the construction as 'go through the Wu Nu star'. He cites the Shih Ching line ' £. Jj£ ^ ' which appears as ' £ ^ |"£.-k* in the Han Shih Ching & (apud. Li Hsia-ting 1965:2453). As the graph I^T appears in the O.B.I, only twice, it is difficult to propose a precise interpretation. For the time being, Lu's interpretation seems to be phonetically and contextually possible and is thus adopted in this thesis. Kuo Mo-jo ^ $L identifies the graph i^tf. as ni (rainbow). The reasons for such an interpretation are: CD "The graph jf^C is the character ni , depicting one male and one female rainbow which have heads at both ends. The shih tien section of Erh Ya ffi] states, 'ti-tung>g$ is called yjl ; ti-tung is a.rainbow; a female rainbow is (called) ch'ieh-^erh ifjC . (Kuo P'u states, '(the term) ch'ieh-erh appears in the Shih Tzu • Ho Yi-hsing states, 'in the Hsi Tu Fu ^ £p flrtjV of the Wen Hsllan , there is the citation from the Shih Tzu which states, 'male and female rainbows are (called) che-i It is probable that the ancients considered f the single rainbow hung ^> J, and the double rainbow ni - 266 -(.2) "There is already a graph hung ^ (a graphic variant of hiing t£X. ) in the O.B.I.,, so this one ( T^CH ) should be ni tfc^ . (T'ung jk_ , p.86b-87) Kuo's first argument is difficult to follow. First of all, from the graph T^rty. itself, there is no way of telling that it is the depiction of two rainbows, one male and one female. The double lines ^ may merely depict breadth of the worm-shape. Secondly, the proposal that 'the ancients considered the single rainbow hung ifex. and the rainbow-couples ni. has no support. Kuo probably gets the idea 'double rainbow' from the term ch'ieh-erh ^ ^ . However, the citation from Shih Tzu J1 ^- shows that the graphs ^ ^ can be written as "ffj ^ . It is very doubtful whether one should take the word §^ literally. ^'T ^ff• )• *s most likely simply a bisyllabic word. Moreover, the sentence ' #x % %3 in If ' shows that J|| ( ^ jj]^ ) can stand for either %X- or '|[ (<%%)• The distinc tion between single and double rainbows, if it ever existed, is not carried in the word fi^ . Concerning the second argument, it should be pointed out that Kuo's interpretation of the graph Ao as equivalent to ( ^X-) can hardly be substantiated either graphically or contextually. Even if rj o does correspond to the word ( $fl. ) , this does not exclude the possibility that there can exist another synonym graph(word) 'tCtf , which corresponds to the word ( <£x ). T'ang's proposal of phonetic-loan is phonologically acceptable. In the Shuo Wen, the word 'top gem of a girdle pendant' is written as J^j *grang which is obviously a variant of the graph 4f$f *grang in the'classics. The phonetic yA J *grang/*gang is phonologically very close to the phonetic TXJ *kang. There is a case where these two phonetics are interchangeable. In the glossary of Wen HsOan , Li Shan ^ J| states, 'the express ion gif[ |^j(.to go up and down) is equivalent to AJ[ . p^^J^^I ^% P%tL% (apud Shuo Wen, p.502). The phonetic-loan relationship between the words /jT, fffjlf is quite reasonable. - 267 -Hstt's interpretation of the I Chou Shu ^[Jji)^ passage is questionable. The I Chou Shu states, " ^£ 4p 1> ^ » *T VP ll » ^^^^ j^- " (underlining and punctuation are HsU's.) From the above passage, there is no obvious reason why ytt fang has to be interpreted as a tribe name. Rather than translating the above sentence as 'T'ai Kung Wang ordered Ytt Fang (a tribe) to come', it may also be translated as 'T'ai Kung Wang is ordered to repel Fang Lai'. (The latter translation is based on K'ung Ch'ao's interpretation which states, ' £ ^^if^jtj'^ ' (I Chou Shu, chap.40^if % , p.9). This might be translated as 'T'ai Kung Wang received the order to chase and repel Ch'ou's ally Fang Lai'.) There is an obvious reason to give preference to this interpretation. The last sentence in the I Chou Shu passage states, 'to announce (one's victory) by presenting the captives and the left ears of the slainj^ <+K %^ '• The action '^ v$ ^}f-Jr ' is always a ceremony performed as a token of victory. 'To order Ytt Fang to come' would not provide a sufficient condition for performing such a ceremony. Moreover, though the verb ming might be interpreted as an active verb, it would be much better in the present context to interpret it as a passive verb 'be ordered'. As the passage focuses on King Wu's defeat of the Shang, other persons mentioned would best be interpreted as officials acting on King Wu's orders. The passage reads, "On the day of ping-wu and the next day ting-wei at the beginning of the first month, the king marched from Chou to attack the Shang King Chou. ... Having arrived at Shang, (King Wu) decapitated one hundred evil officials of the Shang King Chou armed with bows and arrows. T'ai Kung Wang was ordered to repel Fang Lai. At the day ting-mao when Wang arrived, he announced (his victory) by presenting the captives and the left ears of the slain. ... Ltt T'a was ordered to attack YUeh, Hsi and Fang. (On the day) jen-shen, Huang Hsin arrived, he announced (his victory) by presenting the captives and the left ears of the slain. Hou Lai was ordered to attack Mi Chi at Ch'en. (On the day of hsin-ssu, (Hou Lai) arrived, he announced (his victory) by presenting the captives and the left ears of the slain." - 268 -• •• • & K ft #!/ «j^*T^^|.gK,X';f 74 % T *r f k, % ^ t£ 1$. •••• j?ffc ^ # M_ M 7T . £ ? % % % % «A #fc Y# . ^ jfc. f ^ # * f /8t. ^ 2 , ^ (I Chou Shu, Ch.40,p.9) (Notice the parallelism between the actions of T'ai Kung Wang Zv '£ , LU T'a g and Hou Lai f£ ^ .)' The passive usage of ming w^p" without a marker has a parallel in the Chu Shu Chi Nien 1"t it > 'The Chou people attacked the YU Wu barbarians and conquered them. Wang Chi of Chou was mandated to be the Shepherd of Yin.' )H A. /ft £ & iL^y jt^*4 * £ f Ff. (Chu Shu Chi Nien,p.23), The application of word families in deciphering the verbal meaning of the graph(word) zt! and the proposed meaning 'to offer' were both initiated by Takashima in his letter to David Nivison of January 7, 1979 and in Takashima 1980. The author of this thesis has made some changes in describing the word family {£' by sub-dividing it into two groups which have opposite senses and proposed that the meaning 'to offer' may be directly derived from the second group. This term is redefined by Lu Shih-hsien J% % ^ (1971). In the present thesis,the author follows Lu's definition. Before explaining the meaning of 'graphs created through phonetic-lban', a brief remark on the term 'chia-chieh jfi^j^ phonetic-loan'may be desirable. - 269 -In a non-chla-chieh graph, the graphic structure (GS) conveys the meaning of the graph(word). G sound—j——meaning GS-"""" While in a graph used as phonetic-loan, the graphic structure does not show the meaning; the meaning is 'shown' merely by the sound. G sound——-—meaning The term 'graphs created through phonetic-loan' is established on the basis of the above definition. For example, the graph /£ *nawng (SW form ^ ) is a depiction of two hands holding a big shell to plough a field (bronze form , the component 6^ is the original form of the graph ch'en ^ (a huge clam), written as in the O.B.I, and bronze inscriptions. In the Huai Nan Tzu ijj. "J" > there is a sentence '(the ancient people) sharpened a clam to hoey^ ^ V^j^l' ' - (Huai Nan Tzu, ChUan 13 fan lun hsdan >[j »j^|HJ , p.l). (See Kuo Mo-jo 1962:200). The graph *ngwng conveys the meaning of ploughing. But in the hsieh-sheng series of *ngwng, many words have the basic sense 'thickness'. *nj9wng/*niewng/*niawng thick clothes *niawng thick, of liquids; heavy dew *niawng thick liquor (strong liquor) *n9wng_ pus "f!fL *n^£wns/*njfiwng dense, luxuriant (as vegetation) Since the graph ^ *newng_ does not have the sense of 'thickness' in its graphic structure, we may say that jL *ngwng is used 'phonetic-loan' in the graphs *niewng/*niawng/*njawng as a - 270 -y^*ni9wng, *niawng, ~))^ *n9wng and -ffi^ *nj ewng/*niewng. To define chia-chieh-tzu fflX/fy ^ as 'a graph in which the graphic structure does not show the meaning' has the advantage of clarifying the confusing terminology such as hsing-sheng-chien-i jjf j 1^ Jjt'Ulr- (-tne phonetic of a hsieh-sheng graph also conveys the meaning) , hui-i-chien-sheng J^. ^ ^] (one of the components which conveys the meaning also serves as a phonetic) and i-sheng jjJ-> AT, (one of the components which conveys the meaning also serves as a phonetic) which are all used in more or less identical sense. But in fact, these three terms refer to two types of component relationships of a 'combined graph ^"^j? 2r ' ' Type 1. The phonetic component, while serving as a phonetic, ... also conveys the meaning through its graphic structure, e.g. ]^*p'wan/*bwan' (SW form ) According to the Shuo Wen, one of the meanings of this graph is 'half of a body ^j?'jfC/^ '» the component ^ *p_wan graphically means 'separating an ox into two' and thus has the meaning 'half. Combining ^ *pwan and*f$] *njawk/*nj9w into a graph, the meaning 'half of a body' is conveyed. In other words, the semantic combination of these two components can yield the meaning 'half of a body' even if the component *pwan did hot have a pronunciation similar to fflj^ *p'wan/*bwan. The graph can be termed i-sheng-tzu Jff, ^ ^ par excellence. Type 2. Since the theory 'yu wen shuo 3C i*^L> ' has been postulated, some Chinese scholars have started to refer to another type of graph as i-sheng-tzu as well, for example, the graphs t  *niewng/*niawng/*njawng, *niswng, f^^_, *niawng, ,fjtffc_ *nawng and *n jewng/*niawng. They consider that since the graph ^ *nawng, when serving as phonetic, frequently conveys the sense 'thickness', it has the meaning 'thickness'. When using *newng as a phonetic to form a graph, e.g. *ni8wng, the component does not merely serves as phonetic but also has a hui-i (K jg^ function, i.e. j^j(liquor) + ^ (thickness) = strong liquor. - 271 -However, one can easily tell that the relationship between the components of *p'wan/*bwan is different from that of t$^_ *niawng. The meaning (shown in the graphic structure) of the' phonetic *pwan constitutes the meaning of *p 'wan/*bwan while the meaning (shown in the graphic structure) of the^ phonetic fi> *nawng does not. In fact, cases such as GM$C_ *niQwng should be excluded from the i^sheng JfpHT category, since their meaning is not shown by the graphic structure of the phonetic (or more accurately, not shown by the relationship between the graphic structure of each of the two components S and P) but merely shown by the sound of the phonetic. If graphs such as f%f. *niawng can be classified in the category of i-sheng, then all hsieh-sheng graphs should be classified as i-sheng tzu since the phonetic of any of these graphs necessarily conveys the meaning. If we insist on terming graphs such as |^l^L*niawng as i-sheng tzu, the categories of hsieh-sheng tzu g $ % an<^ i-sheng tzu would completely coincide; graphs such as fiffi *p'wan/*bwan would form a sub category within this hsieh-sheng or i-sheng category and a new term would have to be coined for them. On the other hand, if we restrict the category of i-sheng tzu to graphs such *p'wan/*bwan, we should term graphs such as J$\& *niawng and */ *ni9wng as 'graphs created through phonetic-loan'. It is somewhat surprising to find that Shima does not give a separate entry to the graph j( . Shima was probably confused by the practice of symmetricism in inscriptions. Graphs such as ^ , when engraved on the left-hand side of a plastron, often take the form ; when on the right-hand side, the form ',j Thus, every graph ^ is not necessarily representing the word ^ *tsa. A form may be simply a mirror image of the graph(word) *|jw&Y. In Chia ^ 3913 and 1369, we find the following inscriptions: - 272 -These two inscriptions can be nothing else but the mirror images of the very commonly encountered expressions ' ^ (receive abundant assistance)' and ' 32. Jtj-(receive abundant harvest)' respectively. However, there are also examples showing that the graph is not always a mirror image of the graph(word) . ^ is a different word. ting-yu / test / king / make / three / army / right / middle , Jul ^ :mg-yu {J*® Ts'ui & 579 The king should organize (lit. make) three armies, the right, the middle and the left. Moreover, a preliminary review of the graphs (f and ^ listed in the SPrui ^{^(unfortunately, it is not an exhaustive list) shows that the graph(word) j( is most frequently used as a transitive verb while ^ , when not functioning as a sacrificial verb, does not share such a usage. (In the Sorui, there are two inscriptions i.e., Ho 213 and I_ *ZJ 3435, in which ^ seems to be used as a non-sacrificial transitive verb. However, the original rubbings show that these two ^ are, in fact, .) For example, r test / Hsien / indeed / tso / king Hsien will indeed tso the "king. Ping 41 (16) Ha f m f \ U it aatiKm ^ test / Hsien / not / tso / king Hsien will not tso the king. - 273 -tablet / tso / king The (spirits of) tablets will tso the king. if. ## u tablet / not / tso Ping 98 (18) Ping 98 (19) The (spirits of) tables will not tso the king. (The subject(agent) is often a spirit and the object is often the king when they are expressed.) tiU U Mi ft* $*L i chia-shen / crack / Ch'Ueh / test / Ch'Uen / receive / assist Ming flfl 2358 Ch'Ueh will receive assistance. (A non-spiritual person may be the subject(beneficiary) and here functions as an object.) In\.this«thesis, graphs <f and are taken at their surface graphic value, i.e., ^ is read as *tsa and as */jwaY, unless there is evidence suggesting otherwise. 3". ance 8. The graph TH-j- lacks a parallel Shuo Wen form. Graphically, it is composed of ^ ( captive, prisoner) and jH ( axe-like weapon). The identification of ^ with jj^- is well established; the upper portion $ depicts the head and the pigtail (YU Hsing-wu -7 , 1940:27). The bronze forms of this graph are J> & (Chou Fa-kao )% fk] 1974': no. 1362). The lower portion </^> depicts the body with hands bond in the back. It is worth noting that this graph can be written either as <^(Ito 255) orft^ (Ping yh 523), the latter shows the head (and the pigtail) separated from the body. - 274 -(Cf. the graph ,ffj (to dismember) can be written as ( i + (| ) or jflj , the broken tree ^ of the latter shows the effect of the chopping action.) Such a subtle variation strongly motivates us to interpret the graph as a hui-i ^ graph depicting a captive being decapitated with an axe-like weapon. (See also Yao Hsiao-sui 1979a:317.) In the clause ' ;£ Jf^ f '» this graph(word) probably functions as a transitive verb having the meaning 'to decapitate (a captive)'. The reading of this graph(word) awaits further study. 9. The graph ^ has long been a matter of controversy. Kuo Mo-jo interprets it as pao ^ ( ^ wrap) (Ts'ui , no. 1523); T'ang Lan interprets it as a graphic variant of ^ ( ^ pig) (1939:20-23); Yeh Yu-sen J§; % % (1932: Vol.5, p.34), Wang Hsiang % (1925, section on the rituals, p.5) and Lu Shih-hsien ^ ^ (1958:3) interpret it as mao 7% (mao-grass); Tung Tso-pin -f^ ^ interprets it as mao ^ (spear) (1977: Vol.2, p.636); Ting Shan J [Jj interprets it as a graphic variant of hsi jj ( ^ night) (1956:3-9); Tseng I-kung fy&'k interprets as the original form of $ (body) (Cho no.18). Chang Ping-ch'ttan transcribes it as p'i [7L (1972: Vol.3.2, p.18); YU Hsing-wu ^ /ff (1940:1-4) and Li Hsiao-ting ^ ^j^.(1965: 187-189) interpret it as t'un . Among these interpretations, the theories of graphic variant are the most tenuous since there is no explanation why the graph j^- has some pecularities that are not shared by the graphs of which is an alleged graphic variant. Among other graphic analyses, YU Hsing-wu'S theory appears to be most convincing. The graph (word) J^- is written as in the bronzes. The graphic evolution may be as Ytt proposed: O.B.I, common form Ch'ien #<] 7.7.2 T'ieh ffjt 44.4 bronze common form - 275 -Yll proposes two meanings (probably phonetic-loan usages) for this graph: (1) '\J, *dwen bp, *tsj ia//*dwen/*djwan/*tj iwan (silk cloth) 'The king gave Ch'Ueh two bolts of silk cloth.' (2) t^, *tiwan jj^-*t'jiwan (spring) e.g. /^^.^jf, This spring we will receive a (good) harvest. The seven month. But none of these interpretations is applicable to the clause ' %fy % fyj ' " '''ne graPn(word) ^ probably functions as a tribe/place name in this inscription. (YU mentioned that the graph (word)^^ in which Jt appears to be a phonetic may stand for the name of a statelet Tun (Tso Chuan HY 120/^23/3^. ). But he did not suggest that , by itself, may also stand for a tribe/place name.) U -f? H ^* ^ 3-34a keng-hsU / crack / Ch'Ueh / at / T'un / go out Ch'Ueh will go out (to take the field) at T'un. jen-hsU / crack / i-ch'ou / use / Marquis T'un On next i-ch'ou day, we should use (as a sacrificial victim) the Marquis of T'un. ccf. % v ht 4P * « rfl *P M*> (Tso Chuan HY 114/19/3/£ ) - 276 -'The duke of Sung made duke Wen of Choo sacrifice the viscount of Tsang at an alter on the bank of the Suy.' (Legge, p. 177) In the O.B.I., the name of a tribe/place can also stand for the people from that tribe/place. Hi tfU Mm 3* 4*• LL 7128 test / next / chia-wu / use / many / T'un On next chia-wu day, we should use (as sacrificial victims) many T'un-tribesmen. This inscription is comparable to the following inscriptions: •:H|>*p 1s Wffl # *- T? "tfJLVHs{i m 1-44-1 ting-mao / crack / use / captive / to / Brother Chi We should use captives (in a sacrifice) to Brother Chi. h^t r, fl* #•«, ff/ i& 6687 i-wei / crack / use / pig / to / Ancestral Mother I We should use pigs (in a sacrifice) to Ancestral Mother I. The word in the clause ' fyj ' can be interpreted in the same way. 10. It is probable that there are two divinations, the first one extending from ^ to ^X^, and the second from ^ to ^jT . The first inscription goes from top to bottom and then from left to right. The second inscription has only one row and it goes from top to bottom. And the second one is inscribed to the left of the first one. It is extremely unusual, if we are going to consider them as one single inscription, for an inscription to go from left to right and then jump back to the left. Moreover, it is highly unlikely that a - 277 -single inscription would be inscribed spanning both sides of the central line. The author would provide the following two translations for the first and second inscriptions respectively: 'We should perform an exorcism to Ancestral Mother Hsin and Ancestral Mother Kuei.' 'It is a boar (that we should use); it is a sheep (that we should use).' - 278 -11. Concerning the graphs ^i. , ^fi , %L , % , and (j^ , Lo Chen-^u considers the first three graphs to be variant forms of the word C '^X- ) 'male mammal' and the latter three graphs variant forms of the word f ^ ( $b ) 'female mammal' (.1914:27). Lo probably gets this idea from the fact that in the classics, the words £ji£- ( ^* + zt- ) and ( + ) are used to refer to the male and female of any animal be they ox, sheep, or pig. However, Yang Shu-ta points out that Lo's theory cannot be maintained. For example, in Ts'ui 4^ 396, there is an inscription fx ^1 Q ^j-(fa if both and merely mean 'male', the inscription would be translated as 'male male white pig', which is very strange. Therefore,^ Yang proposes that fx , ^p- and ^i-refer to 'bull, male sheep and boar' respectively; while f^ , and ^~J_ refer to 'female ox, female sheep and sow' (1974 r Vol..1, p. 2-3) At first sight, these graphs look like hsieh-sheng f0 ^ graphs; having f ( ), T ( $ )»]^(^<)as signifies and ± (..£•) t ^ ( -CJ ) as phonetics. However, if this is the case, ( >M ), Tx ( •)» }± CJ«- ) on the one hand and fj ( $tL> ), ^ ( )» ^ ( ) on tne otner hand would have very similar or even identical pronunciations (there are also some other graphs having 'deer' or 'dog' as one component and J— ( £- ) or f) ( (j ) as the other component). Since these words often appear in identical linguistic environments, i.e., as direct objects (pati.'ent's'J sacrificial items), it would be very difficult to avoid confusion. Another possibility is that they are hui-i ^ graphs.' f (ox) + JL (male) fx. (male ox) 'C|f (sheep) + _L (male) ^ (male sheep), etc. This solution has one advantage, it avoids the difficulty caused by the discrepancy between the graphic structure and the pronunciation of the graph fx. ( ). The Shuo Wen defines this graph as 'male mammals, deriving from (the signific) ^ and has jjt- as a phonetic' (SW, p.51). Since the reconstruction of fyi- is *mwjw and j^. is - 279 -*t*af/*daY, the graph j£- could hardly function as a phonetic in the graph J^f" . Some scholars, such as Tuan YU-ts'ai, have suggested that the graph should be the graph *dz±st instead (ibid.)* While in terms of the finals, *dzist (. 2^%^ ) might possibly be linked phonetically to the graph *mwaw ( ), the initials *dz- and *m- are totally unrelated. If we consider tfcf-a hui-i graph, we would not face such phonological difficulties. If these graphs are hui-i graphs, they may have completely unrelated pronunciations. In this connection, there are some modern graphs which may be the later forms (constructed by the hsieh-sheng method) of these words. For example, *dak (male ox) *ter (male sheep) $i *kra£ (boar) *rier (castrated sheep or female sheep) *prwaT (sow) However, there are some other graphs which suggest that graphs such as and fx may not be hui-i graphs. For example, there is a graph ^41 (Ch'ien 1.33.7). It is highly unlikely that this graph is a hui-i graph which has a meaning of 'male and female 'ox'. Bearing in mind that ho-t'i-tzu such as JJ ( . Jj + ZJ ) "jJS. ZJ are frequently encountered in the O.B.I., it would not be unreasonable to hypothesize that graphs such as ^j/_L and fi_ are ho-t'i-tzu. Just as JJ is sometimes engraved as j , i.e., as two separate graphs, is sometimes engraved as f) Sj^ • Whether the ho-t'i-tzu and their equi valents are pronounced in the same way or not is unclear to us. There may be a method similar to fan-chieh fj] in pronouncing the ho-t'i-tzu, e.g. is pronounced as *bjw£jf *bjier/*bjien + Jj^ *ngjwe Y But considering there is a graph , the fan-chieh way does not seem to be possible. - 280 -12. There is no obvious reason why we must regard these inscriptions as forming only one single divination. Judging from the original positioning of the graphs, the passage from "J" to tj may be one divination (the expression ^ goes with this divination) and from ^ }} to ^ ^ maY form three others. The author would suggest the follow translations: 'On chia-shen day, (diviner so and so) tested,. "(If we chi-sacrifice to Hsiao I, there will be no harm". This was applied.' 'It is a female (one that we should use).' 'It is a male (one that we should use).' 'We should not yu-offer.' - 281 -13. In saying 'we do not encounter such particles in the O.B.I.', it is meant that there are no graphs(words) which appear exclusively in sentence final position and can be graphically or phonologically related to words such as yeh Jf^, and ;jL. ^ 14. Both Kuan Hsieh-ch'u (1953:38) and Ch'en Meng-chia (1956:128) take hu as an interrogative particle in the inscription 15. This is not an intact bone (or plastron) piece, whether these six graphs form one single inscription is difficult to ascertain. 16. This appears to be an anti-structuralistic approach. But such an approach has its practical usefulness. For the rationale of this approach, see Wang Li's 7V 1946:21-24. 17. The question of verb deletion has been treated in Takashima 1973:163-170. Takashima's analysis of the verb deleted sentences is adopted in this thesis. For the underlying rationale of verb deletion, see chapter two, p.98 of this thesis. - 282 -NOTES FOR CHAPTER TWO There is a great controversy concerning the meaning of the graph(word)^ and a generally accepted interpretation is lacking. (See Li Hsiao-ting 1965:1953-1957.) Nevertheless, most scholars agree that this graph(word) stands for a temporal term. For the time being, the author follows Keightley's functional translation 'season' as a tentative interpretation (1978:78, fn.85). YU Hsing-wu interprets this graph(word) as ni 'a kind of wild rice ; and in this particular context, it functions as a verb 'to plant the ni-rice' (1979:251-252). YU's interpretation is graphically and contextually acceptable. On the same piece I_ Z* 3212, there is the following inscrip tion which may be considered as related to the one cited: m una f'% Pu / plough / at / Tzu / receive / harvest If Pu ploughs at Tzu, (he or we) will receive (abundant) harvest. The graphs^/) and ^% appear in similar syntactic position, thus suggesting that JT£. may also be an agricultural activity. The translation of the sentence ' & /f_«#*f is comparable to an English sentence 'don't come to work tomorrow and I won't pay you'. According to Seiler, this sentence has the meaning 'it is not that I don't want you to come to work; quite the contrary'(1971:81) . In his comments to an early version of the present thesis, Pulleyblank noted as follows which may be relevant to the present observation: "Sentences in which an explanatory clause in the form of an additional noun predication follows the main clause are quite normal in classical Chinese. The difference is that they are regularly marked by the final particle yeh and sometimes also have an introductory particle, especially kai . E.g. (Mencius HY 40/5B/3) - 283 -'There might only be coarse rice and soup of vegetables, but he always ate his fill, not daring to do otherwise.' (Legge, p.378) The English conjunction 'for', ... , is often appropriate in such cases. It could be used in the above Mencius example; 'for he did not dare do otherwise.' ... The only difference in the O.B. example is that there is no yeh Jo2* and no substitute of any kind." Arguing from other evidence, Takashima suggests the word conveys the sense 'specifically' (.1973:389-392). Such an interpretation supports our theory that the focal point of this group of divinations lies on the number of sacrificial victims. Wang Yin-chi jj ^ in his Ching Chuan Shih Tzu f^j^^ t^j does not mention the verbal usage, i.e., to go, of yjl ^ (1974:36-41). James Legge, in his translation of Shang Shu, sometimes interprets yjl as -fj 'to go' (1960:366-367); but he also sometimes interprets ytt ^ as a particle in his translation of Shih Ching; for instance in ^ ^ Jjjj (1960:6). Also, Ch'U Wan-li v£ interprets ytt ^ in some Shang Shu passages as 'to go' (1972:72), but not in the Shih Ching lines such as •^J^-j ^(1967:3). Chou Fa-kao takes the ytt J? in J>,Jq)fjj}r as a verb prefix (1962:251). Yoshikawa Koj iro -j£ llj ^ 5^ , in his translation of the Shih Ching, sometimes interprets yjl as if 'to go' (1969: Vol.3, p.52) and sometimes interprets it as a chu-tzu ^> ^ (ibid.399). In his 'Glosses to the Book of Odes', no.575, Karlgren states, "In the Shi, Ytt 8 is very often an (empty particle), and it is often difficult to determine where it really means 'to go' or simply is a particle: even in the common phrase 10 (.This young lady goes to her new home) there has been much discussion on this point. But there are many good ex. in which ytt 8 clearly is the verb 'to go': Shu: Ta Kao II (I respectfully go with them); ibid. 12 (I will go and attack Yin); ode 194, phr. 13 (You say: to go and take office ), etc." - 284 -8. The expression tsai fei ^ in modern Chinese, when functioning as an independent sentence, means that at the moment when the sentence is uttered, the action 'flying' is in process. If 'the yellow birds are flying', they cannot be 'collected on the thickly growing trees'. 9. In the Shih Ching 177 (Liu Ytieh ), there occurs what appears to be an anomalous sentence ' ^ Z^X ^ ' * 'Our accoutrements were compteted (sic), And we march thirty le_ (every day).' (Legge, p.282) Legge's translation is based on the Cheng Chien which states, "after finishing preparing our military accoutrements, the king was going to dispatch the army and advised 'Proceed thirty li each day, then you may station.'" (SSCCS, chUan 10.2, p.4):. . Cheng interprets yjl ^ as 'to proceed', a usage taking a.numeral. complement which lacks parallels anywhere in the classics. Furthermore, his addition of '^C^.^3 : | i%5> ' is simply an unconvincing attempt to wrest a meaning from these two lines. Karlgren, on the other hand, translates the two lines as: 'our clothes were prepared in our (areas of) 30 li (our homesteads). (1950:120) This interpretation is more acceptable. 10. There are cases where a verbalized noun follows the word yjl -j~ in the Shih Ching, for example, ~.- L & 1 46- (HY 31/154/1) 'In the days of our third month, they take their ploughs in hand.' (Legge, p.226) 'In the days of the third month, we go to plough.' (Karlgren, 1950:97) - 285 -11. In the Tso Chuan, there is one questionable sentence * ffijiffc ~f ' (HY 281/ JjL:14/^3) where yjl f appears to be a full verb. But to interpret yd ^ as either 'be at' or 'go* does not make much sense in the context. The passage reads, .H- % n ;* «i *s*..**-^ (HY 281/^14/3) The historiographers make their records; the blind make their poems; the musicians recite their satires and remonstrances; the great officers admonish and instruct, and inferior officers report to these what they hear; the common people utter their complaints; the merchants (display their wares) in the market places; the hundred artificers exhibit their skilful contrivances. (Legge, p.466-467) Tu Ytl fa ^ glosses lU fjjt as a verb V^- & 'to display' (Legge's translation is probably based on Tu's gloss). But as Takezoe Mitsuhiro Hi X,v>|) argues, shang 10 j?j should be taken as one single constituent. Takezoe suggests that a verb* 5^' 'to discuss' has been omitted here 0)% & X> - I£ % ^ )• This is quite possible in view of the Han Shu reflex of this passage ' /"r., A. /t lh_ * ^ ' (tne common people utter their complaints in the streets; the itinerant merchants discuss in the market places) cited by Takezoe (1961: Vol.3, 14, p.58). 12. In the chapter 'To Shih' of the Shang Shu itb ^ ^ , there is one sentence ' )j) YT) -f % g • £ )£] /§ $ (SSTC 34^) where yU appears to be a full verb 'to go'. However, it is also possible to interpret it as a 'preposition' introducing the locative complement hsin i^ lo_ ' J:,^ (^ ' to the verb kao ' ' . 13. The word 1 ^k. (bone and bronze form: (j ) can function as a full verb in the O.B.I, and bronze inscriptions: - 286 -ing-yu / crack / Ya / Pi / lead / mass ting-yu cross / at / Lu / approve If Ya and Pi lead the mass to cross (the river) at Lu, (there will be) approval. Ts'ui jf^- 1178 it ft ^# c^U /li^^) test / king I order / Pi / lead / mass / pierce attack / Chao /receive / assistance If the king orders Pi to lead the mass to pierce and attack the Chao (Statelet), we will receive assistance. Jin /v__ 2525 not / lead / our / chariot / trap / at / danger (difficulty) (You) did not lead our chariots to be trapped in n difficulties (dangers). (San Tai 5-^ , chuan 9, p.48, Pu Kuei J\> f£_ ) Po Mao Fu / lead / Yin / eight / army / attack / east / I (San Tai 3- yl\ » chuan 9, p.11, Po Mao Fu Kuei |r Po Mao Fu led eight Yin armies to attack the eastern I (barbarian) A check on the HY indices of the Mo Tzu ^- , Chuang Tzu ^ , Hsun Tzu ^ j~ and the Shih San Ching f z shows that there is only one case in the chapter 'Lo Kao' of the Shang Shu jfy ^| i>fe g.^where wang takes a locative complement. ^ ^ (SSTC 33Q239) - 287 -In the O.B.I., wang can take a locative complement, frequently in the form ' 1 .) 15. A sentence such as ' J. ^ \$? 1 may be translated as 'the king will go to hunt' where 'go' is the main verb. But such a translation may not pre cisely reflect the grammatical structure of the Chinese sentence. Chao Yuen-ren states, 'Verbal expressions in series (V-V series) form an intermediate type between coordinate and subordinate constructions, but nearer the latter than the former.' (1968:325) That the first verb is subordinate to the second verb is particularly true in the case of ' jfi, + V, since the content of the predicate is carried almost totally by the second verb. Even if we delete the first verb wang in ' ^ ^3^® ' > the information conveyed is virtually the same. But if we delete the second verb t'ien ^J? , i.e.j'J. ^x- , the meaning of the sentence is greatly altered. 16. The graph j< is graphically equivalent to the Shuo Wen form £ *kuewng which HsU Shen defines as 'to raise clasped hands ^ J$L- ' (SW p.104). Both Ch'U Wan-li ^%o\ (1961:182) and Li Hsiao-ting \ \ \ (1965:0781) interpret it as ^ *kj ewng (offer, present). Yang Shu-ta %ty > on the other hand, interprets it as 7$. *tang (to raise) and takes it as a phonetic-loan of \%*k\ *tia//*ti9ng/*di3ng (to recruit) "(1974:24). While Yang's interpretation fits the context very well, i.e., j/i^ to raise troops, to recruit people, fc\ and (SW form ) can hardly be related either graphically or phonologically. Ch'U and Li's interpretation appears to be preferable. And in fact, in certain contexts,'to offer people' would have a similar meaning as 'to raise (recruit) people', i.e., to offer people (to the king) = to raise (recruit) people (for the king). - 288 -17. According to Chang Ping-ch'Uan, Ping 521 (13) reads sp, tfg (ik-r) • The last graph "ty^l , however, is rather far apart from the graph . The graph ^/j^ may belong to another inscription, i.e., ^ ^ 'to inspect at ^ - ^h' (Ping 521 (.15)). The correct reading of these two inscriptions is debatable. 18. There are quite a few inscriptions where the practice of deletion has left only the preposition and the complement (beneficiary). For example, r \ f. *£>! to / great Ancestral Mother Chi Ping 457 (5) To great Ancestral Mother Chi f f jfj & ^ St should not I to I great Ancestral Mother Chi 5St Should not to great Ancestral Mother Chi 457 (6) In this pair, the verb and its direct object are obviously deleted. See the following pair which appears on the same plastron: ^ ?F M H ft* exorcise / disaster / to / Ancestral Mother Chi We should exorcise the disaster (the exorcism being directed) to Ancestral Mother Chi. 457 (3) should not / exorcise / disaster / to / Ancestral Mother Chi 457 (4) We should not exorcise the disaster (the exorcism being directed) to Ancestral Mother Chi. Another example: ( should not / to / Nan Keng Should not to Nan Keng. L exorcise / disaster / Nan Keng Ping 47 (.10) 47 (9) - 289 -We should exorcise the disaster (the exorcism being directed) to Nan Keng. See also Ping 206 (7) (8). Needless to say, we cannot interpret the word yU in the structure 'yU + ancestral name' as a full verb 'to go'. x The fact that the word yU *}* in \ uj" and ^)-f'[,u) is open to two interpretations, i.e., a full verb or a preposition, is due to the semantic ambiguity of the word ho v"J . But in view of the predominant usage of ho v»J , i.e., the River god, it is better to interpret "f £<"3 and ^7) -f '/v) in the same way as ^ ( M^X- ) j?, and ij? -jr l^tJc • . Ping 443 is a more intact reconstructed version of Ping 290. Chang Ping-ch'Uan, in transcribing the O.B.I, into modern graphs, suggests that the graphs of these two inscriptions should be read in this order, i.e., i * -f \T) and -ft") . However, it is rather difficult to make sense out of these two inscriptions when ordered as Chang's has them. Shima Kunio, on the other hand, gives the following order ' ^ iHj \H ' and ' ^? -f -|L kA. ' (Sorui, p.21-.l) Such an ordering makes much better sense. The author tentatively proposes the following translations: '(We should pray and dance) to the River (god)., (for) we will have ensuing (rain)' and 'we should not (pray and dance) to the River(god), (for we will not perhaps have) ensuing (rain).' Undeniably, there are many words supplied to these two inscriptions. The rationale for supplying these words are: i. On the reverse side of this plastron, i.e., Ping 442, the following inscriptions are found: Ping (6) j 0j) ? <fy ^ gf- 41 tf9 We should present a dance, (for) we will have rain. Ping (.7) On next ting-mao day, we should not present (a dance), (for) we will have no rain. - 290 -The River god is one of the dieties most frequently associated in the O.B.I, with the rain-soliciting sacrifice. Thus, to associate ^ 'present, a dance' (a rite often performed to solicit rain) with 'River god' is very natural. ii. As there are a lot of chisel holes on Ping 443, it is quite possible that many graphs have been obliterated. Thus, the present author supplies a graph "fc, before , and (<K? after i.e., ' iZj jfcjl iL&l ' and ' ^ \7) <£. Ik (^'. Though there is no positive evidence to support such a supplement, it fits the general O.B.I, pattern very well. (Cf. the inscription 'we will have ensuing rain' on Ping 442. 20. The graph is not shown in the rubbing, but Chang Ping-ch'Uan supplies it in his kao shih (1959: Vol.1.1, p.17). Since Chang is the one who had access to the original plastron pieces, he may well have been able to discern some graphs that were not reproduced in the rubbings. 21. Ping 510 is not an intact plastron, the section right below the graphs ' |j J[ /Jjjj ' has^not been discovered. Shima Kunio puts some dots under the graph ( ), showing that he considers there are some graphs missing (Sorui, p.241.4). 22. In the Shih Ching, the word fang y) is occasionally used as a sacrificial verb. For example, V*A fa \XA (HY 52/211/2) We sacrifice to (the spirit of) the land, and to (those) of) the four quarters. (Legge, p.377) $t i (HY 52/212/4) They will come and offer pure sacrifices to the Spirits of the four quarters. (Legge, p.382) - 291 -7y fa ^ % • (HY 70/258/6) I was not late (in sacrificing) to (the Spirits) of the four quarters of the land. (Legge, p.532) The theme of odes 211 (Pu t'ien j|f |2 ) and 212 (Ta t'ien 7^ 27 ) is agricultural activities, while ode 258 (Yun han ^J^ ) is about drought, thus also related to agriculture. It is natural to arrive at the conclusion that the fang-sacrifice is a rite to invoke a good harvest or success in agriculture. Such an interpretation of the verb fang 'fj can be applied to the O.B.I.: i. When the effect (consequence) the Shang people wish to bring about by offering a fang-sacrifice (which frequently co-occurs with other sacrifices) is expressed, it is invariably related to agriculture, for example, ... wu / crack / fang-sacrifice / ti-sacrifice / three / pig V* fiC $ tr If yu-offer / dog / dismember-sacrifice I to I earth(god) penned-sheep / invoke / rain I_ 77L 40 We should fang-sacrifice and ti-sacrifice three pigs, yu-offer one dog and dismember-sacrifice a penned-sheep (to present) to the earth(god) in order to invoke rain. fang-sacrifice / burn-sacrifice / hui / keng • 14* 247 yu-sacrif ice / have / big / rain ±1^ l*1 rn In performing a fang-sacrifice and a burn-sacrifice, if, on a keng day, we yu-sacrifice, we will have heavy rain. - 292 -ii. The fang-sacrifice is frequently associated with sacrifices directed to the earth god (t'u j^, , she ^i, ) who is also closely related to agriculture. In addition to the inscription I_ ffi 40, there are Ho 211, Ts'un 7^. 1.595 and I_ Z-> 5272 where both the fang-sacrifice and sacrifices directed to the earth god appear. For example, burn-sacrifice / to / earth(god) / penned-sheep / fang-f 4 sacrifice / ti-sacrifice Ho /n 211 We should burn-sacrifice to the earth(god)a penned-sheep, and perform a fang and a ti-sacrifice. •t'% fl 4i burn-sacrifice / to / earth(god) / fang-sacrifice ti-sacrifice Ts'un 1:.595 We should burn-sacrifice to the earth(god), and perform a fang and a ti-sacrifice. In fact, the co-practice ' \*A^^- ^A 7) * recorded in the Shih Ching (HY 52/211/2) may be a tradition inherited from the Shang rituals. (The quasi-identity between j£- *f_aT/*da/ and \i- *dj_iajf has been conclusively illustrated by Wang Kuo-wei J, )^) ^jjwho cites the following evidence: i. The Mao Chuan glosses the term chung t'u j£yi-5-as 'a large she & ' . ii. The Shih Ching line ' fiJL Y^ 'is cited as ' %5 pC YC ' in the Shih Chi. iii. Ho Hsiu fat glosses the word t'u ^ as she . (.1968: Vol.6, p. 1959) Since a sacrifice directed to the earth (god) t'u 5— (she ) can - 293 -be called a she-sacrifice, it seems likely that the name fang 7j" comes from the fact that such a rite is directed to the spirit of quarters f]. In the O.B.I., the spirit of quarters 7J , just as the fang-sacrifice, is related invariably to agriculture. $:* H U fn _ perhaps / invoke / harvest / to / quarter(god) / receive harvest Nan Ming 425 If we perhaps invoke for a good harvest to the quarters(god), we will receive a good harvest. & £ %% - to f* f % TsJ^ui ^ 1545 perhaps / pacify / rain I to I quarter(god) Perhaps (we should perform a rite) to the quarters(god) to stop the rain. Also, among some fifty verbal occurrences of the word fang & (sacrifice), there is not a single inscription in which fang 3 takes an indirect object. (There is one inscription tT ^ -5 ' 854) where the name Father I appears to be the indirect object of fang 4, However, since part of this inscription is obliterated, we cannot exclude the possibility that there may be a verb in front of the word yjl , e-g« * tT A. <—> > <J (••• burn-sacrifice to Father I and perform a fang-sacrifice). The absence of the indirect object can be explained by the hypothesis that a fang-sacrifice is invariably directed to the quarters(god), so that semantically it has already incorporated in it the indirect object, and there is not the necessity to express the indirect object in the surface structure. Although we do not know exactly how the presumed fang-sacrifice was practised, inscription I 40 $)L , ^ i\ , xjf'p-f =^ ( ... to fang-sacrifice and ti-sacrifice three pigs, yu-offer a dog and dismember-sacrifice to the earth(god)^ a penned-sheep)' suggests that fang may be a ritual-sacrifice incorporating the killing of victims. In other words, the particular way of killing victim in a ritual performed to the fang-spirits is called a fang-sacrifice. If so, fang - 294 " is a transitive verb which takes sacrificial victims as direct objects. Applying this analysis to Ho ^ 211 cited on p.114, we may interpret, as reflected in the translations, the word nU ^ (woman) as the direct object of the fang-sacrifice. As another possibility, an interpretation of 77 ^ as: 'antiposited object + verb' (i.e. to (the spirit of) the quarters, we ti-sacrifice) may be derived from YU Hsing-wu (1979:187) and Shima (1958:203). In corroboration of such an interpretation, there are inscriptions which incorporate the construction ' f Tj to ti-sacrifice to (the spirits of) the quarters' (e.g. Pa ^ 9, see also Ping 216 (1), (2), (3), (4)). However, it is difficult to explain why, if ^7 is the object of the verb , 77 Is most often placed in front of 'rj7 (30 occurrences) while the structure ^ ^ jj ox rfy f %. / / &r / it / ff (altogether 5 occurrences) occurs so rarely. Among the 30 occurrences of Tj ifp / tyti^, we never encounter the structure * -f if ^ I /fV f Tj • (Cf. ^ff £-$%7 "f to exorcise the king's misfortune (the excorcismbeing directed) to Ch'iang Chia (I £j 842)^ -J ^ f ^ ftjtj to Ch'iang Chia we should perform an exorcism against the king's misfortune (I 2, 842).) Why then is the presumed antiposited object jj never introduced by the co-verb yU J~ ? Also, there are the following inscriptions: | ? ^ ^ , H o £ 211 We should burn-sacrifice to the Earth(god), fang ti. Tj % 3- % , 2 *L >r 1 i % , fcfa 1^ ™ We should fang ti three pigs, yu-offer a dog and dismember-sacrifice to the Earth(god) a penned-sheep to invoke for rain. Notice that the indirect object(beneficiary) j£. is placed after the verb ||_ / /f j7 ,i.e., ^/^f7 + f + > but tne presumed OB 7? is not placed after the verb w , i.e., * ^ + ( f ) + Tj '. If both X- and f] are indirect objects(beneficiaries), why do they not show parallel structure? Until the above questions can be satisfactorily - 295 -answered, the author has reservations in accepting the interpretation that 77 is an antiposited object (indirect object) in the structure jj . (See also Serruys 1974:72). In the Shuo Wen, there is a graph (word) fang J/f', a variant of 4^) . which HsU Shen defines as "a sacrifice performed inside the gate, whereby the (spirits) of the ancestors wander about" n ft ft.fcfaftrMtftl (sw, p.4) The definition of this word given by HsU Shen is well supported by the classics. The Li Ch'i 4|L $B chapter of the Li Chi ^ t£j states, "Inside the great temple .... prepare a sacrifice at the hall, and have the fang-ritual outside (the hall)." *- n ••" %&%f * Pl 4 ft (SSCCS, chUan 24,p.l2) The great temple ( A $H ) is a place where the ancestors are worshipped, the fang-ritual is presumably directed to ancestors in a way similar to that of HsU Shen's description. Also, the graph fang appears in the Shih Ching: ^ f frf (HY 51/209/2) "The priest sacrifices inside the temple gates." (Legge, p.369) The Mao Chuan fa \^ glosses fang jfijj as 'inside the gate OZJ (SSCCS, chUan 13.2, p.7). Thus, although the graph(word) ^jj appears to be a later form of in the O.B.I., the different recipients of these two ritual-sacrifices suggest that they may be two different ritual-sacrifices. On the other hand, in the O.B.I, there is another ritual-sacrificial verb ^ (or ) of which the recipients are almost exclusively the Shang ancestors. HsU Shen states, - 296 -[I is a container/case; it is a pictograph, read as "ft ~~1 is the chou wen form of d'"m (SW, p.641) T'ang Lan proposes that the ^) -ritual in the O.B.I, is equivalent to the 'jffj ritual-sacrifice in the classics (apud Li Hsiao-ting 1965:3820) Although it is difficult to draw comparisons between the ritual-sacrifices in the O.B.I, and those in the classics since we do not know the exact character of these ritual-sacrifices, T'ang's proposal seems to be possible in terms of the similarity of recipients and the pronunciations. If fang in the classics is a ritual-sacrifice derived from one of the O.B.I.'s, J , instead of /j"*, appears to be a more likely origin. - 297 -NOTES FOR CHAPTER THREE 1. The word •*> (to harm) can take an object, e.g., (our envoy will perhaps harm the statelet) (Ping 69 (1)). But it is not a controllable verb. (1) It never takes the negative wu 4? . In a great majority of cases, the negative is fu , in a very few cases, it is wu . (2) Although it might appear to be quite possible, sentences such as + X1 (to order/to call upon (so and so) to harm (X)) in fact never appear. In composite sentences, this verb frequently appears in the main clause to represent the state or consequence of a certain action: fif ^ 1 K h h'l H§ A^ 0e hsin-ch'ou / crack / Ch'Ueh / test / this / day Prince (?) / perhaps / trap / Chi Fang Fou / harm Today Prince will perhaps trap Chi Fang Fou (the Fou people of the Chi Statelet), he will harm them. Ping 302 (1) chi-mao / test / order / Ch'a / lead / multitude •f # I %t j \ :tack / Lung / harm v attack If we order Ch'a to lead the mass to attack the Lung (Statelet), we will harm them. Or: We ordered Ch'a to lead the mass to attack the Lung(Statelet), we will harm them. K'u ^' 1001 If we use semantic terms, may be classified as an 'success' verb which implies that the successful completion of a certain act is - 298 -intrinsic to the meaning of the verb. Also belonging to this category are verbs such as ^ 'capture' and ^ 'capture (by net)'. (The notion of 'success verb' has been drawn to the attention of the author by Takashima.) The word /" ( *j( ) *tsa rarely functions as a sacrificial-verb (only twice in the Ping Pien), thus it is very difficult to decipher its meaning. The word family of f' show that it has a meaning 'to hurt, to eliminate'; and there is a word *dzra which means 'to cut trees'. Although solid evidence is lacking, this tentative interpretation, i.e., 'to cut', seems to fit the context. One might surmise that ^ is a mirror image of the graph ( ^1 , 5^ ) 'yu-offering', one of the most common sacrificial verbs in the O.B.I. However, the inscription Ping 96 (11) is engraved on the right-hand side of the plastron, a position which would lead the Shang inscriber to engrave even the word ^ *tsa in the form ^ . It is unlikely that, in such a position, the presumed word ^ *(fjwaT would be engraved as ^ . The word jo_ To- functions either as a transitive verb or a stative verb. For example, (1) Transitive usage: ... Ch'Ueh / test / tablet / approve / king I 8091 The (spirits of) the tablets will approve the king. test / Hsien / approve / king Hsien will approve the king. HsU t?5.7. (2) Stative usage: lit ft ; / king / please The king will be pleased. test / king / please 1 7771 - 299 -king / not / please I o 7771 •The king will not be pleased. Thus, the clause ' "% ' can be either interpreted as 'ti-god will approve (X)' where X stands for an unexpressed object; or 'ti-god will be pleased.' In the inscriptions in question, the author prefers the second interpretation since it does not assume an unexpressed object. It should be pointed out that this particular pattern, in collections of rubbings other than the Ping Pien, is not limited to divination about the building of a settlement ( ). For example: king / should not / hsiang 7 go out / tablets / please Cho*y| 165 If the king does not specifically(?) go out, (the spirits of) the tablets will be pleased. Or: Let the king not specifically(?) go out, and (thev,spirits of) the tablets will be pleased. The graph ^<^is obviously a graphic variant of as these two graphs occur in highly similar environments. T>t* 11- £ f m it' WJ, keng-yin / crack / Cheng / test / king / perhaps teach / not / encounter ... Ning -»p 3.95 The king will perhaps teach (the people), he will not encounter (rain)• As to the graph , most scholars interpret it as»( ' ) 'to teach, to enlighten' (Lo Chen-yll, 1914 : Vol..2 ,.p. 61').'.Phonologically and graphically, this interpretation is acceptable. A hypothetical - 300 -graphic transformation of can be outlined as follows: ft r„ 0 ^ ^ Ning %. 3.95 ^ Jin 60 —* Ching ^ 4836 * Ta YU Ting % ^_ ijfe SW form C] *graw (The reconstruction is based on the word x > the presumed phonetic in this graph.) i^t. *graw (. ^ *grawk, this graph is recorded in the SW as a variant of The graph ( ^ ) is obviously polyphonic. In the HsUeh Chi Chapter of the Li Chi /jtjf , there is a sentence ' £p , ^ ^ teaching and learning are equally (important).' The glossary states, 'the first word is (pronounced) as hu hsiao fan, i^f? , Jo- £fl %~ fa^ ' (SSCCS, ChUan 36, p.2). Such a fan-chieh points to a pronunciation similar to that of *kraw (to teach) for which the Kuang Yun gives the reading ku hsiao chieh "jj^ P7f (YU Nai-yung 1974:415) (The variation between these two readings, i.e., |lp ( 2f$t__) *kraw and >£p *grawk, seems to involve a morphological process.) Such an interpretation fits the O.B.I, texts very well. ting-ssu / crack / Ch'Ueh / test / king / teach ft ftf f? If , * mass / attack / to / Mou Statelet / receive / abundant assistance Ping 22 (5) The king should teach the mass to attack the Mou Statelet, (for) he will receive abundant assistance. (For the interpretation of "/j , see Yu Hsing-wu 1979:16.) - 301 -cf. it |_ & ^ ^ (Mo Tzu HY 35/21/10) 'Then (the sage kings) taught the people martial activities.' The interpretation of the graph td is very problematic. First of tall, it is difficult to tell whether the upper element f represents a knife 79 or a body.outline If it is the former, this graph -77 k-may be transcribed as {?. » the alternative yields fip in k ai-shu. (Shima lists this graph under both 'radicals' in Sorui: 356 and 6.) Here, we tentatively transcribe it as fa. • In either case, ^ functions as a proper name in the great majority of inscriptions. Such a nominal usage simply does not fit the inscription in question. Nevertheless, in Hsu ^.S^ 2.24.5, there is a partially obliterated inscription: y i- ri ft ^ * ff hsin-hai / crack / Ch'ueh / test / yu-offer / to A - if m Mieh I ... I ^ I two I ... I dog / announce-sacrifice flye / ... / ox We should yu-offer to Mieh two ... dogs, and announce-sacrif ice five ... oxen. If the obliterated graph after the graph ^ (two) is a word representing a sacrificial victim (e.g. *fyi. pig) , a highly likely possibility, then t/ can be interpreted as a method of sacrificing victims.. However, the rare verbal usage of this graph (word) makes it difficult to substantiate further such an interpretation. The word wang seems to have a particular ritual usage in addition to its common meaning 'to go'. test / this /day / wang / to / (Shang) Chia Nan Ch'eng 17 - 302 -To day, we should wang-sacrifice(?) to (Shang) Chia. (See also Ts'un 2.187) ping-yin / crack / Ch'Ueh / test fw f ft f * ' & * Ancestral Mother Keng / yu-offer / woman wang / two / ox / next / keng-wu / use (To) Ancestral Mother Keng, we should yu-offer a woman and wang-sacrifice(?) two oxen; this sacrifice(?) will be applied on the next keng-wu day. I 341 kuei-ssu / crack / wang / ffy (?) / bring / rain We should wang-sacrifice (?) to to bring about rain. Nan Ming 429 (There are various divergent opinions about the interpretation of the graph (see Li Hsiao-ting 1965:2839-2846). While a precise and universally-accepted interpretation of this graph still eludes us, there is no doubt that it stands for the name of a spirit. For example, Ts'ui $f 15 *4C ^ ~f f% . $ ^ 1 'if we will e> . - - , - -perhaps invoke rain from we should-..burn-sacr if ice', nine penned-sheep.) , It seems natural to interpret sentences such as as to go (in a magical sense) to Shang Chia —> to approach Shang Chia', since the possibility of communion between spirits and men was a fundamental element in ancient Chinese religion. A well-known expression of this belief is found in the lines: i £ l*^ ?4 fa ^ ^ (Shih Ching HY 58/235/1) 'King Wan ascends and descends, on the left and the right of God' (Legge, p.428). However, there are also cases where takes victim-objects: - 303 -hsin-hai / crack / ti-sacrifice / wang-sacrifice (?) - - T ? ** one / sheep / female pig We should ti-sacrifice and wang-sacrifice (?) a sheep and a female pig. I_ Z» 5327 Or: In performing a ti-sacrifice and going to (the spirits), we should (USE) a sheep and a pig. chi-mao / crack / # (?) / wang-sacrif ice (?) /-.ox K'u 1124 We should wang-sacrifice a pied ox(?). Or: In approaching (the spirits), we should (USE) a pied ox(?) (As the graph occurs only once in Shima's Sorui, it is impossible to provide a conclusive interpretation. The author suspects that it is a graphic variant (or simply a scribal error) of the ho-ti-tzu "^T^t ( }$J ~f — a Pied ox) • For example, ... / crack / Ta I ... /'ting-mao sui-cut-sacrifice / female ox We should sui-cut-sacrif ice a female ox. Ching Jp, 3218 . (See also HsU 1.30.6, 1.30.7; Hou Shang ^'_b3.5.) There is not sufficient material to determine whether this ritual-sacrifice wang ^i- is a method of sacrificing animals or of approaching the spirits. If we adopt the former interpretation, wang 7*£ might be treated as a causative verb * to cause to go —> to send'. For example, .--^ i< 'to send two oxen (to Ancestral Mother Keng)' (I ^ 341, cited on p. 302 ). (This is an interpretation suggested by Takashima during a private consultation.) In either case, it seems fairly clear that wang is related to rain soliciting: - 304 -test / wang-s acr if ice (,? ) / to / C?) / have ensure / rain I_ 19 If we wang-sacrifice (?) to (?), we will have ensuing rain. Or: If (we) approach , we will have ensuing rain. call upon / wang-sacrif ice (,?) / to / River (god) have / ensue / rain I_ c* 1465 ... if we call upon so and so to wang-sacrifice (?) to the River(god), we will have ensuing rain. Or: ... if we call upon so and so to approach the River (god), we will have ensuing rain. (See also T'ieh 70.3) (Ytl Hsing-wu interprets 4^ *gjwang as a phonetic-loan of *njang (1979:154-156).) Ping 71(6) and Ping 73(6) are members of a ch'eng-t'ao ( )• Thus, these two inscriptions are counted as only one occurrence. Ping 73(6) reads ' N ,*f $ ™ ' ( | , fc}). The graph *f (Ad ) is obviously an scribal error for the graph "^f ( ). ' ^ ^/ ' ^°eS n0t ma^e much sense. Nor can we interpret r^ as an abbreviated kan-chih ^ combination, since, in such a case, it is generally the earthly member (i.e. i^L. )» not the celestial member (i.e. ^ ) that is abbreviated. That is, Chia-hsu ^ becomes Chia ^ , not *hsu ty^ . - 305 -There are only two exceptions among 173 inscriptions where the word y_U occurs. have no / perhaps / rain Ping 499 (4) We will perhaps have no rain. Jft ViE ¥(« -ifc keng-yin / have / ensue / rain Ping 158 (3) On the keng-yin day, we will have ensuing rain. (The rubbing of Ping 158 (.3) is very unclear; here we rely on Chang Ping-ch'Uan's reconstruction.) As to Ping 158 (.3) , we should point out that on the reverse side of this plastron, i.e., Ping 157, there are the inscriptions ' Jtjjk, , y.' kk (tf£) ' (If we burn at stake a woman (of Tsai) , we will have ensuing rain) and ((If we burn at stake a woman of) Tsai, we will perhaps not have ensuing rain.) Thus, it is reasonable to interpret the sentence ' '<|? vj* Jk ' as an ^ abbreviated version of a hypothetical sentence '* //\ f^*) •' That is, rather than merely indicating the possibility of raining at a future time, the sentence 1 4i Ak ' expresses the intended result of a rain soliciting ritual-sacrifice. Thus Ping 158(3')! ceases to be a real exception. It should also be pointed out that Ping 499 is not an intact plastron and we may be lacking a wider context for this, apparent exception to the above generalization. This generalization is based on a review of the ming-tz'u /*J? jffj^f in the Ping Pien. The following is a list of the inscriptions incor porating the word 'rain - 306 -A. B. 157 (1) C2) 158 (3) 199 (9) (10) 223 (5) (6) C9) 280 CD (2) 340 CD 431 (6) 442 (6) (7) (9) 463 C2) C3) 469 CD 499 (4) the reverse side of 157 see 442 (.6) (7) a more fully reconstructed plastron rubbing of 223 C J, ) mi ^ cjt )K&7 or C-Jt ) W «fr I ^ < J£ ^ ch'eng t'ao ^ J^" (1) 3 (15) (16) 32 (9) 43 (1) 63 (3) (4) (5) (.6) C7) (8) 71 (6) 73 (6) 84 CD 87 (6) (7) 93 (1) (2) 104 CD 116 (3) (.10) CH) 149 (.21) (.22) 153 (7) (.8) (.13) (14) 202 (9) C10) (11) (12) 211 (14) (.15) same, as 306 220 (2) (.3) 235 (1) C2) (5) (6) - 307 -242 (2) 252 (2) 260 (2) (3) 263 (9) (10) 266 (2) 304 (1) (2) 306 (8) (9) same as 211 333 (7) 342 (21) (22) 343 (8) (9) 346 (2) (3) 346 (4) (5) (46) (47) 379 (1) 392 (5) 393 (8) 422 (3) (4) 425 (6) (7) 433 (1) (2) (3) (4) 440 (2) 447 (8) (9) 474 (1) 508 (2) 515 (14) 519 (5) (6) 520 (3) 521 (6) (7) 522 (1) (2) (3) 523 (5) (6) 525 (4) 527 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) 528 (1) (2) 529 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) 531 (13) 532 (1) 534 (4) (5) 565 (1) 622 (2) (3) - 308 -(2) . fl x % x, c ) / % o#r- ) i-fe or §7 x, c J~ ) 1^7 / ^-^r «fo (X stands for a temporal term) 151 CD C2) C3) (4) 155 C5) C6) 255 (1) (2) 381 (7) (8) 390 (3) (4) 454 (.1) (2) (3) CA) 525 (3) (3) cJr ) >t(fo/ 2^% 149 (11) (12) 387 (7) (8) 529 (D (2) (3) C4) (5) (6) (7) (8) C9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (4) ' (_A ) fa I # ( ) ' appears in the first clause of a composite sentence 203 (22) (23) 211 (16) (17) 306 (6) (7) 335 (.17) •c. % R_fc> 280 (1) (2) 340 (1) same as 306 same as 211 - 309 -10. The nature of the referent of the word ssu 7i awaits further invest igation. However, it is obvious that it refers to a supernatural fir Ho Vz 318 Ho & 318 being. ( In <** fr* ft it go / inspect / Shu / SsuC?) / please p ^ ... go to inspect at Shu,.'Ssu'(?)will be pleased. v Shu / Ssu(?) / not / please ... (go to inspect at) Shu, Ssu(?) will not be pleased. Xt ^ j h ^* |i kuei-wei / crack / Ch'Ueh / test / king / capture (?) Ssu(?) / approve Ch'ien ty) 5.30.5 (If) the king captures(animals), Ssu(?) will be pleased. These inscriptions are very similar to the following ones: king / make / settlement / ti-god / please i-go * If the king makes a settlement, ti-god will be pleased. Hou Hsia test / Chia / being at / here / tablet / please Chia is here, the (spirits) of the tablets will be pleased. Ping 276 (2) 11. That the structure yu yU ~fty represents the intended result of certain rain-soliciting-ritual-sacrifice does not mean that yu 7^ conveys the sense of 'to obtain intentionally'. Since ' misfortune', (in the structure 7$ I rfty \$h ) apparently is not something the Shang want to obtain intentionally. In fact, the word yu in both the structures 7$ ifc) and "7$ ?)z / ^k^^^o) functions simply as a transitive verb 'to have, to get'. Also whether something is an - 310 -intended effect or not lies .on. the semantic character of the word/ words following yu_ "7^ , rather than on v_u 7^ itself. The eight divinatory groups provided by Serruys are mainly taken from the first two hundred rubbings of the Ping Pien. The author has searched the remaining four hundred and twelve rubbings (as some of them are duplicates, the total is six hundred and twelve rubbings, not six hundred and thirty-two) and finds the following groups of inscriptions where ch'i appear in both positive and negative sentences 227 (3), (4) main verbs: (will perhaps bring . . .) Jf; Cwill perhaps not bring ...) subject: i|c (?) 259 (24), (25) main verbs : /*^- (will perhaps break in .. . ) (will perhaps not break in ...) subject unexpressed. 261 (8), (9) A u , main verbs: >p -J; T (will not perhaps call upon ...) «J (will perhaps call upon ...) subj ect: Ch'Ueh M. 306 (12), (13) main verbs: (will perhaps take) J|_^f $K Cwill not perhaps take) subject: Chia j|-330 (10), (11) main verbs: ^ ( ) (will perhaps not come) S-§Z 5n Cwill perhaps come) subject: Cha ^ 436 CD, (2) main verbs: ^ (will perhaps have the occasion) tx-^T (will not perhaps have the occasion) subject: (?) i - 311 -527 (.3), CA), C7), C8), C9) , (10) main verbs: ^ [$} (will perhaps rain) f^} Cwill perhaps not rain) Notice that all the verbs are marked by the *p— type negatives and hence uncontrollable. QPing 436 (2) is the only apparent exception where the negative of ^ is expressed not by a particle but by the main verb *mjwang . However, it is also uncontrollable.) 13. The interpretation of the subordinate clauses in I 2/ 960, I_ Z 4626, Ho 210 and the like as simultaneous-successive rather than conditional is based on semantic considerations. The possibility that the Shang would make the following kind of divinations seems rather remote. *If (so and so) has an ailing eye, he will recover. I Z 960 k ^ & JM* £ ^ I -L 4626 *If (so and so) has an ailing tooth, it is not (because of) Father I's harm. Or: *If (so and so) has an ailing tooth, it is not Father I who causes the harm. f *If (so and so) has an ailing eye, it will not be prolonged. Ho_ 210 *If (so and so) has an ailing eye, it will perhaps ^be prolonged. Moreover, upon checking Ping Pien, it is discovered that, with only one exception, wherever the following structures appear, independent sentences divining about whether someone is sick never appear on the same plastron. - 311 -527 (3), (4), (7), C8), (9), (10) main verbs: 1^7 (will perhaps rain) (will perhaps not rain) Notice that all the verbs are marked by the *p- type negatives and hence uncontrollable. (Ping 436 (2) is the only apparent exception where the negative of Lj^ is expressed not by a particle but by the main verb *mjwang X— • However, it is also uncontrollable.) 13. The interpretation of the subordinate clauses in I ^ 960, IC ZLJ 4626, Ho /gT 210 and the like as simultaneous-successive rather than conditional is based on semantic considerations. The possibility that the Shang would make the following kind of divinations seems rather remote. *If (so and so) has an ailing eye, he will recover. I 960 | V fit Ml, I i 4626 *If (so and so) has, an ailing.tooth, ;it is not (because of) Father I's harm. Or: *If (so and so) has an ailing tooth, it is not Father I who causes the harm. *If (so and so) has an ailing eye, it will not be prolonged. Ho_ ^ 210 < *If (so and so) has an ailing eye, it will perhaps be prolonged. Moreover, upon checking Ping Pien, it is discovered that, with only one exception, wherever the following structures appear, independent sentences divining about whether someone is sick never appear on the same plastron. - 312 -1 (NP) + ^ + >Vk + (body part), main clause' e.g. ^.ff , frf""^ f£:T (Ping. 508 (5)) Someone has an ailing trunk, we should perform an exorcism to Ancestor Ting. or '(NP) + + (body part), main clause' e-g- & % (£injj 239 (6)) Someone is suffering from an ailing tooth, it is not (because of) there is harm (from the spirits). That is, a sentence such as (Prince YU, perhaps, has an ailment) (Ch'ien ft>\ 5.44.2) does not co-occur with composite sentences similar to the two cited above. Such a phenomenon suggests that sentences such as ' \£ A % , */» ?1& T ' and l£y $ % ' refer to a sickness already incurred; and thus, it would be redundant to divine about whether the individual in question is sick or not. The only exception to this generalization is found on Ping 96. i-ssu / crack / Ch'Ueh / test / have / sick / trunk favorable Ping 96 (21) < (So and so) has an ailing trunk; he/she will recover (lit. it will be favorable). i-ssu / crack / Ch'Ueh / test / have / sick / trunk not / perhaps / favorable (So and so) has an ailing trunk; he/she will perhaps not recover (lit. it will not be favorable). 96 (20) test / king / perhaps / sick / bone The king, perhaps, is suffering from ailing bones. - 313 -Unfortunately, the subject of the clauses ' y A H is not expressed, we cannot tell whether this pair refer to the king or to another person who had an ailing trunk. Following is a list of the inscriptions incorporating the word chi -vf!fo in the Ping Pien: A. chi iffc followed by a main clause 12 (8) (9) 14 (.8) (.9) 16 (8) (9) 18 (8) (9) 20 (8) (9) 27 (3) 29 (2) 31 (2) 51 (8) C9) 52 (14) 88 (1) (2) 96 (20) (21) 239 (5) (6) 258 (10) 356 (3) (4) 473 (3) (4) 508 (5) 541 (3) (.4) r ch'eng t'ao } ch'eng t'ao B. chi not followed by a main clause 33 (.6) 96 (5) 106 (.18) (.19) 175 (1) (.2) 190.'CD (2) 212 (1) (2) 243 (7) - 314 -251 CD (2) (.7) (.8) 285 (.4) 316 C16) 334 (.1) (2) C4) (5) 502 (.13) (14) 508 CD C8) 549 (1) (2) As the graph occurs only twice on the Sorui, its meaning is difficult to determine. Since it follows the co-verb yU -^p in this inscription, it may well represent the name of a place. However, the structure of this graph also suggests that it may be a kind of game. In the Shuo Wen, there is a graph jH^ *kiar (a variant of ) which is defined as 'a big mi-deer ^VJK & ' (SW:475). If the component i^, *kjiwar in the graph functions as a phonetic, then the pronunciations of the graphs /i& and \M '7\ would be sufficiently similar to suggest their identity. And we may interpret yU as a conjunction, i.e., 'and'. See Han Yao-lung 1973:10. In these two tui-chen pairs, the protases of the negativecounterparts are not expressed, however, we may ass.ume a hypothetical full form •'•* £/-tf(&$: SJ/1L' (if the king/we perhaps chase rhinoceroses/game, the king/we will perhaps not capture (them).) It is highly unlikely that these two negative counterparts are ^ derived from the structure '* (the king/we should not chase rhinoceroses/game, (for) wewill'perhaps not capture (them).) Since such a 'cause and effect' structure would be abbreviated as '* £ / 4$ 5fc / /^fy ' Cthe king/we should not chase rhinoceroses/game); instead of being abbreviated as •'* eLj CCfor) the king/we will perhaps not capture (them).) The clause ' ' > if being treated as a subordinate explanatory clause 'for ...', does not constitute a meaningful charging statement. (See Chapter Two, p. 88-91 for a discussion - 315 -of the practice of ellipsis.) J 7 L 16. The graph -jlr appears only five times in the Sorui, and in only two of these cases are the inscriptions intact. It is difficult to provide a. .reliable decipherment. This graph is composed of a kuo like weapon and hsing which often signifies military activities. (See Chapter Four, fn.23). Thus, we may hypothesize that this graph (word) conveys a meaning similar to that of fa h.^ - However, there is no way to tell whether it is a hsieh-sheng or a hui-i character, not to mention the subtle distinctions in meaning which may exist between this graph (word) and the word fa . 17. Ch' en Meng—chia transcribes the graph as and reads it as hsiang (a personal name)(1936:502). Chang Ping-ch'Uan transcribes it as 1^.(1962:336). Ch'en's transcription is followed by Ikeda (1964:2.35). Lu Shih-hsien transcribes it as ^ (apud Ikeda 1964:2.35) • It is often difficult to speculate on exactly what kind of animal a pictograph represents by merely looking at the outline of the graph, and as Takashima has noted, none of these scholars' opinions seems to be based on solid evidence (1973:359). 1 Although it is difficult to tell what animal Ml represents, the graphic structure of J± suggests that it may have a meaning similar to that of 'to trap' (an interpretation proposed by Serruys, see Takashima 1973:35). Cf. jjj^j (a graph depicting an animal in a pit, generally interpreted as 'to trap'). The component is probably an alternative from of ( J (a pit, a cave). f and (—J are found interchangeable in the following graphs: A'- iA Ching \ 4 ' 4- $ ' Li Ha 78 • i%\ % : u}6 Ch'ien jft 5.28.6 ' ^ ^ * & Hou Shang $% £ 29.10 'jj $>. * & Mj : $j Ch'ien jpj 7.40.1 ' HsU 5.2.4 ' 316 -To interpret as 'to trap' fits the context very well. This graph(word) appears in linguistic environments highly similar to that of 'to capture' and ^ 'to capture (by net)'. f test / many / prince / chase / game / trap If many princes chase game(?), they will (successfully) trap (them). Ping 43 (3) ping-shen / crack / Ch'Ueh / test / we / perhaps $ * $ 41 chase / game(?) / capture If we perhaps chase game(?), we will capture them. I_ 2. 6728 kuei-yu / crack / chase / (?) / capture (by net) If we chase ^? (?), we will capture them (by net). Choffi|232 test / not / perhaps / trap Ping 43 (2) (If many princes chase game), they will (successfully) trap (them). Prince / ffi (?) / not / perhaps / capture Prince ^ will perhaps not capture (animals) . I_ 2, 5956. not / perhaps / capture(by net) (So and so) will perhaps not capture (animals)(by net). Chia y 37,55 V Q may also be a 'success' verb. - 317 -18. Philological notes on T'ung jf|f_ 735 (1) ^£5 fifej'• this expression occurs only three times all of which are yen-tz 'u /JJ^- (verifying statements). Kuo Mo-jo's inter pretation 'just as prognosticated' -%af>fj TJL (T'ung ;jg_ , p.158) seems to fit the context nicely and thus has been adopted. (2) l5 ' one of the interpretations of rJ7 given by the Erh Ya is chih 'to take charge, to govern, to preside' (HY 3/1B/29). Such a meaning can be applied to the O.B.I, expression ' jij £ j^- ' (Chia ^ 3338) 'to take charge of the king's affair; and to the Shih Ching line ' % ^ ^ (HY 25/121/1.2.3) The king's affairs are not being (well) presided over' (the author's trans lation) . The word ku seems also to be related semantically to ku J^f 'hard, solid; to strengthen', (see Keightley 1979:25). (.3) ' the interpretation — (precipitously) is proposed by Derek Herforth who cites a parallel semantic development in Latin: pareceps 'headlong' > steep(cliff) > (falling) quickly See also the Shih Ching line ' <ff'] i_ 1$ ' (MI 54/220/4) 'With their caps on one side and like to fall off (Legge, p.399). 19. Although there is no strong evidence in support of this hypothesis, the author would point out that, according to the records (yen-tz'u fyf^ ) in the O.B.I., the number of animals captured in the Period I is often greater than that of the later periods. There are cases where the Shang captured hundreds of animals in one hunting set out in the Period I, e.g., Ping 284 (4) (372 animals), Ping 87 (5) (451 animals), and Hou Hsia //fy "p. 1.4 (287 animals), 41.12 (348 animals). Hunting on such a great scale is not found in the later periods. This fact seems to suggest that hunting was an important, and presumably, a practical activity in the Period I. See Huang Jan-wei 1964-65 'Yin - 318 -Wang T'ien Lieh K'ao KB and Ch'en P'an 1965 'Ku She Hui T'ien Shou Yu Chi Ssu Chih Kuan Hsi* \d- ^ Kit ffi ^ - 319 -NOTES FOR CHAPTER FOUR Since Lo Chen-yU who transcribed the graph ''ffj as J/tJ) 'wine' (1914:Vol.2, p.25), most scholars have interpreted it as a ritual which makes use of wine (Li Hsiao-ting 1965:4399, Shima 1958:268). Thus it is often equated to the word kuan which the Shuo Wen defines as 'the ritual libation' (SW, p.6). However, there is certain doubt about this transcription, and thus the interpretation. Although Lo Chen-yU did not elaborate on the rationale of his transcription, it appears that he takes the three strokes as the component equivalent to the water radical ^ (SW form 'f, ). If this is the case, Lo's transcription is justified. But as Yeh Ytt-sen has pointed out, the three strokes should be interpreted as yung £ ( fj^ ) a graph also appearing in the O.B.I. (1932: chUan 1, p.47) as the name of an R-S activity. Upon checking the inscriptions, it is found that the word (or the element) 'water' is rarely engraved in the form of three slant parallel strokes but rather as scattered dots. For example, /K '// T'iehfS' 14-3 ;g p-l Ch'ien 7.13.3 it* ft) Nan Ming 432 Ch'ien 6.43.1 Chia ^ 3337 The last two graphs have not been deciphered. Nevertheless, there is no doubt that they depict the pouring of liquid from one container to another'. Notice how the drops of liquid are engraved. In fact, there is another graph which, graphically and contextually, is a much more likely candidate for the primitive form of the word 'to perform a libation', namely, ^ ' (Snlma 1958:266). This graph has many variant forms, but the basic structure is a vessel (wine-pot)-- 320 -like object ty; Ping 217 (2)). It may take additional elements and form graphs such as (Ping 217 (.1)), Tj( (HsU 1.38.6) . As pointed out by Shima (1958:267) there are two striking facts about this graph: (1) In some inscriptions, there are dots(presumably representing drops of liquid) engraved close to the month of the wine pot, : i.e., I* Ping 217 (2) fr\ Ming 1391 ^ Ch'ien 4.33.2 Such a realistic and vivid depiction strongly suggests that this graph(word) represents the wine pouring. Taking into consideration that this wine pouring ritual is performed in front of a tablet ' ~J~ ' (as suggested by the first two graphs), it seems most reasonable to interpret it as 'to perform a libation'. (2) The graph(word) f$ never takes animal-victims as direct objects. In the rare cases where it takes a direct object, the direct object is 'aromatic wine' ^ , e.g. M| «4t f$*?? test / aromatic wine / perhaps / perform libation K/u 1253 (It is) the aromatic wine which we should perhaps pour. This inscription coincides with the Mao's glossary (of the Shih Ching), (Kuan means to pour aromatic wine' (SSCCS, chtian 16, p.ll). Although this graph does not seem to have a phonetic element, the pictographic structure and the object it takes can be used as evidence to justify the interpretation 'kuan ' (to perform libation). - 321 -Of course, that there is a graph (word) ^J^T / representing the word 'to perforin libation' does not necessarily exclude the possibility that the graph (word) may also represent a meaning identical or slightly different from that of kuan ( ). (The graphs £f|} and •Jf-jj^ cannot be graphic variants since they can both appear in the same inscription, e.g., Ts'ui 393 and Liu Ch'ing 191 (Sorui 155.4). The major reason to refute such an interpretation is that it does not fit the context: n,ph r fl* £ ping-wu / crack / Pin / test / chop-sacrifice ' w T % ^% ft & , * *i eight / sheep / and / yu-sacrifice / thirty ox / eight / month Jin A. 83 (1) We should chop-sacrifice eight sheep and yu-sacrifice thirty oxen. At the eighth month. X*V"A n M* $*) tf kuei-yu / crack / Cheng / test / next / chia-hsU cut-sacrifice / ten / sheep / i-hai / yu-sacrifice ten / ox I 404 On the next chia-hsU day, we should cut-sacrifice ten sheep; and on the i-hai day, we should yu-sacrifice ten oxen. The above inscriptions are cited by Kaizuka who casts doubt on the traditional interpretation ' — libation' (1960:167 and 551). As Kaizuka argues, the parallelism between the words ^ ("ll^f ) 'chop-sacrifice', $ (. ) 'cut-sacrifice' and ($)^ 'yu-sacrifice' suggests that yu is also a method of sacrificing victims. (Kaizuka transcribes the graph as jj^Cibid.' 167) and he does not provide definite interpretations for the two graphs and $ .) (For the interpretation °f ^ ) 'chop-sacrifice', see YU Hsing-wu 1941:29; and also p;208 . - 322 -of this chapter.) In the case of $ ), there are also cases illustrating that it is a method of sacrificing victims: invoke / harvest / to / Mu(?) / cut-sacrifice sheep / burn-sacrifice / penned-sheep dismember-sacrif ice / ox _I Yf\ 153 In invoking for a (good) harvest to Mu(?), we should cut-sacrifice a sheep, burn-sacrifice a penned-sheep and dismember-sacrifice an ox. (Notice the parallelism between the words ( ), ^ and ty^ .) ping-ch'en / crack / h (?) / test / perhaps ) J} WW) \«L °T ^ f * cut-sacrifice / Father Ting / three / penned-sneep • I_ 725 We should perhaps cut-sacrifice to Father Ting three penned-sheep. The above inscriptions demonstrate that Kaizuka's suspicion is justified. One might argue, the structure should be analysed as 'we should perform a libation (with/by USING) thirty/ten oxen', a structure parallel to ' jfc "T* ^ ' (Ts'ui ^ 680) 'we should invoke (with/by USING) ten oxen.' (Such an interpretation is fully discussed in the body of this chapter, p.201 - 208). However, there is a major difference between the linguistic behaviour of $fG) and ^\ (and also J( j5 , ^ ). In the case of , an OB extremely rarely appears after the OV, i.e. ' >f\ + OV (. ^ ) + OB'; instead, we find the pattern ' ifi + ( ) + OB + OV' (see the statistics on p. 200). On the contrary, the pattern ' + OV + (. -f ) + OB' is not at all rare, for example, - 323 -2fa+Jr*if+%+'£L HsU f'| 2.21.4 tfa + 5^ + l^'L Hsulf 1.3.3 (Cited on p.196-197). (See also Nan Ming P$ 99, Ch'ien "^fj 8.13.4, Nan Ming 126, Ts'ui ^ 528, Ho /fe" 289 (two occurrences).) In its linguistic behaviour, resembles^ and ^ , i.e., the OV functions as an OP instead of an 01. Kuo Mo-jo interprets as a phonet ic loan for . Concerning the inscription '^. ^(tjj , ft\ . . .' , Kuo states, " >(*> is loan for#, . The Ta Tsung Po Chapter of the Chou Kuan states, 'pile up wood to burn-sacrifice to worship^ Ssu Chung and Ssu Ming' . The words and co-occur, (a structure) similar to this inscription. The Ssu Tien $tL>-$x Chapter of the Feng Su T'ung states, ' ^jf^ means to pile up wood to burn-sacrifice." (T'ung jfflf.-, p.3) For the time being, Kuo's interpretation is adopted as a working hypothesis. Many scholars have worked on the decipherment of the graph (tj and yet none of them has arrived at a generally accepted conclusion. (For a summarized review of these various interpretations, see Serruys 1974:96 and Mickel 1976:126-131.) The graph has variant forms ^ and /^j . That they are graphic variants can be demonstrated by the following inscriptions: Ping 98,(12) «2£ t M Ping 98 (14) ' $ *r 3- hi - 324 -a a I 2 3441 liSS. 32 (17) ' t B M' P±ng_ 83 (8) ' $ £j M ' ^£83 C7) ' £ The graph jij looks so similar to the ox scapula that it is hardly deniable that it is a pictograph depicting a scapula (Chin Hsiang-heng 1959:3.35). See the attached picture taken from Tung Tso-pin 1933:173. The last two forms J^j , £jf are"graphically closely related to the commonly encountered graph 7$) ('misfortune' ( ^TJ , , , /^|>, ) which also points to the shape of a scapula (Kuo Mo-jo 1934:10; Li Hsiao-ting 1965:1493). In addition to being graphically highly similar, all these graphs convey a bad sense: H £ ^ # * ci t $ test / wei / Ancestral Mother Chi / inflict king / ^ Ping 462 (1) It is Ancestral Mother Chi who inflicts ^ upon the king. - 325 -ti-god / perhaps / make / king / misfortune Ti-god will perhaps cause misfortunes upon the king (lit. make the king misfortunes). I_ ~Z> 4861 test / king / gjf / wei / noxious-influence The king ^MJ -ed, it is (because there is) noxious influence. Ho 286 5af*. U ?It H| ti 1ft* chi-wei / crack / Ch'Ueh / test / king / dream noxious-influence / wei / misfortune Ho ^ 301 The king dreamt about noxious-influence, it (means) misfortune. Also, as Mickel pointed out (1976:132) the graph ^ frequently appears after the word iff () 'to exorcise* (1.2, 2095, 1099, 7572). Obviously is something undesirable, otherwise the Shang people would not try to exorcise it. For the time being, we tentatively translate it as 'misfortune' (nominal usage) or 'harm,kill' (verbal usage). The fact that the graphs /jnf , , /jf and jj[ , , /^j are graphically so similar and they both share a bad sense leaves open the possibility that they are mere variant forms of the same word. Nevertheless, there appears to be a complementary distribution in terms of their, linguistic behaviour. _The graphs jjr^ , ^ are extremely rarely used in the expressions ^*fi (to cause misfortune) and %$Jj (to have misfortune) while these are the expressions where we most commonly find the graphs Jj\ , and . On the other hand, the graphs /jj » ^ » are rarely used in.the construction ' £p + personal name + X' (to exorcise so and so's X)(where X stands for the graphs in - 326 -question) while ^fj frequently appears in this context. Moreover, members of these two groups of graphs sometimes co-occur in a single inscription: test / Hu (?) Statelet / go out / not / wei 0/ we / being at / Huo The Hu(?) Statelet has gone out (i.e. taken the field), it will not (cause) misfortune. We (made the divination) at Huo. I iX. 172 f 1 % $ Hi ^ t & $ & chia-shen / test / Hsing Statelet / come / wei H / I / being at / Huo Ping 319 (1) The Hsing Statelet is coming, it (means) misfortune. I (made the divination) at Huo. These two inscriptions show that the graphs ^rr| , J[ and jjj are not freely interchangeable. Graphically and semantically these two groups of graphs are so close and yet they are kept distinct. Such a phenomenon leads us to hypothesize that these two groups of graphs (words) have a common origin and are in the course of diversification. This diversification seems to have a grammatical character: The graph(word) (or ^ .... etc.) is used as a nominal term in a great majority of inscriptions. For example, f Mi3!* B *• fl ** test / we / have no / misfortune We will have no misfortunes. Ping 3 (12) test / we / perhaps / have / misfortune We will perhaps have misfortunes. Ping 3 (11) - 327 -(Sentences similar to the ones above number in the hundreds. Although Shima has not provided an exhaustive list of the inscriptions incorp orating the expression ^jJ £T) / £j £TJ , among some two hundred occurrences of these expressions, we fail to find the graph jJ ( Jj ).) On the other hand, although the graph ^ ( Jj ) sometimes also show a nominal usage as in the expression (fj % ^'(17^ 842) 'to exorcise the king's misfortune', it exhibits a verbal character rarely shared by the graph (word) jjj (- /jjf ) (observations based on the data collected in the Sorui): test / wei / Father I /harm / king It is Father I who harms (lit. to 'cause' misfortune ^ upon the king). Ho ^ 286 H| VM> %n }L ffii ii. i. test / riot / wei / Father I / harm / king Ho ^ 286 It is not Father I who harms (lit. causes misfortune upon) the king. burn-sacrifice I to I east / three / small / penned-sheep ^ -sacrifice / dog7^ Ts'ui ffi 462 We should burn-sacrifice three small penned-sheep to (the spirit of) the east and /^J -sacrifice a dog. chia-shen / crack / Pin / test / burn-sacrifice tT ft ~£ — ^ T# J to / east / three / pig / three / sheep / /gj -sacrifice dog / dismember-sacrifice / yellow / ox We should burn-sacrifice to (the spirit) of the east three pigs, three sheep, |£j -sacrifice a dog and dismember-sacrif ice a yellow ox. HsUJ^l .53.1 - 328 -Moreover, as mentioned above, the construction '* £f/ ^ ' (have misfortune) does not occur. (Shima lists two inscriptions, i.e., I_ 8289 and Chia ^ 3510 which incorporate the construction ^£ ftfj . However, upon checking the rubbings, it is found out that in J. "Zi 8289, it is the graph j£. (."£,), instead of , that occurs. Chia 3510 is blurred, so we cannot ascertain whether it is the graph ^ or the graph JJ which actually occurs there. Taking this into consideration, sentences such as ' fjjj , /ft & % ' (I Z- 4033, 7495) and ';£-/§, ^(= )' (I 2, 3441, K^u 7^. 1975) may be interpreted as 'the king (is experiencing) misfortune, it is (because) of harm (from the spirits)' and 'the king (is experiencing) misfortune, he will be all right' respectively. That is, ^ffj is interpreted here as a stative verb. Concerning the expression ' [j\ 7\ ' (Ts 'ui ^jp 462, HsU „tf 1.53.1), since £7] shows a coordinative relationship with and , we may propose that it is also a type B verb (see p. 219—221). If the pronunciation of ^rj is similar to that of ^ ( *krwe/ , *gwa ) , we may tentatively interpret it as ^'J *k_^wejf ^ which is a common sacrificial verb in the classics, for example, (SSCCS, Li Chi, chUan 43, p.13) After the temple is completed, (the yung-j en y$,k^ ) ^/J-kills a sheep (and let) the blood flow in front of it. V\ $ ' CI Ching, HY 33/54/_t ) A young man 'I.J-kills a sheep, (but) there is no blood. (SSCCS, I Li, chUan 47, p.7) ... to go to the position outside the entrance of the temple.... the official ssu-ma ^'ij -kills a sheep. (See also SSCCS, Li Chi, chUan 17, p.21; chUan 28, p.7) - 329 -Mickel, by combining the face value of the graph ^jrj (i.e., bone) and the suggestion of misfortune conveyed by it, proposes that £?rj represents a kind of bone illness, probably rheumatoid arthritis (1976:131). Such an interpretation seems to be also possible. For the time being, a definitive decipherment of this graph is elusive. There does not seem to be any established term to represent the common characteristic of the OGs (e.g. £- jj ' (the activity of) the T'u Statelet1, 'misfortune', if-'harvest') in relation to type A verbs (e.g. -%'to announce', '-r^'to exorcise', rf<'to invoke'). For the time being, we call these objects 'object-goal (OG)' represent ing the goal or cause of an R-S activity. An alternative term is 'object-patient', since, for example, the misfortune ( ) can be thought of as something being exorcised ( ^fj7 ) However, as the following passages in the body of this chapter illustrate object-victims can also be object-patients . (OP), i.e., *V ( % , fff , *H ) + ( f ) + OB + OV-OP' or 'VB ( J , ^ , yf ) + OV-OP + ( f ) + OB' Thus, in order to avoid confusion, the term 'object-patient' is not used to refer to an object of this kind. Other possible terms, such as 'object-content' (e.g. the activity of the T'u Statelet is the content of the announcement '3= jj ) and 'object-unaffected' (e.g. the activity of the T'u Statelet is not affected by the announcement jf ) have been considered. However, none of these terms seems appropriate in the case of ^-f7 Z.fyuj'to exorcise the king's misfortunes' since is neither the content nor is it something unaffected. A term which abstracts the common characteristics of these objects is still lacking. - 330 -Since there are no better choices, the term OG is still employed. Admittedly, it does not, in all cases, appropriately reflect the relationship between the object and its respective verb. Nevertheless, it has its own merits. That is, in relation to the structure 'V^ + OV' representing an 'instrumental' activity, an object of this kind enjoys the status of 'goal', i.e., it is because of such an object (matter) that the OV (of the type B verb) is used (sacrificed). 4. In classical Chinese, when the direct object is placed in front of the indirect object, the co-verb ytt -f I is used to introduce the indirect object. Otherwise, ytt y / jf^ Is not present. This is a rule drawn to the attention of the author by Pulleyblank in his comments to the draft of this thesis. 5. In Sorui 54.1, Shima transcribes I 8810 as • ^ \ f& = ^ %\ ( T £ f~ , £p S k^-M, )> i-e., ' ?P P + OV + OB', a structure which would be an exception to our generalization. However, upon checking the original copy, it is discovered that the following reading may equally be defended, based on the ambiguous arrangement of the characters in the inscription (see the attached xerox copy): ting-ssu / crack / yu-sacrifice / west / exorcise' Ancestral Mother Keng / three / penned-oxen We should yu-sacrifice, (and then to/at the) west we should perform an exorcism to Ancestral Mother Keng (with) three penned-oxen. - 331 -Although such reading of the inscription is somewhat unusual, it is not totally impossible. If it is accepted, I 8810 ceases to be an exception. 6. The exceptional inscription of ping-tzu / crack / invoke / ox / to / Ancestor Keng We should invoke (with an ox) to Ancestor Keng. I 2. 8406 4 7. The exceptional inscription of Ch'Ueh / test / Hu(?) Statelet / return"'/ follow attack / Pu / king / announce / to / Ancestor I perhaps / punitively attack / seek / assistance ten / month The Hu(?) Statelet came back and subsequently attacked Pu, the king should announce to Ancestor I that we will perhaps punitively attack (it) to ask for assistance. (Divined at the tenth month.) Nan Ming 79 8. The exceptional inscription of TK invoke / to / River(god) / harvest / have / rain If we invoke to the River(god) for a (good) harvest, we will have rain. Ts 'ui 834 - 332 -9. In the classics, the object (which we term as OG in the O.B.I.) can also be introduced by the co-verb i_ , e.g. f ft 1t *A % tf * %i t % (Tso Chuan HY 104/ ft 11/1 £. ) ... the marquis of Tsin sent an announcement to Loo of the disorder attempted to be raised by P'e Ch'ing. (Legge, p.158) X- «A ^ %fa % f # (Tso Chuan HY 112/*§-16/ ?1f 3) The king sent word to Ts'e of the troubles still raised by the Jung. (Legge, p.171) However, the co-verb i_ is not an obligatory element and can be omitted without altering the cognitive meaning, for example, 1t U fi% & fe f % (Tso Chuan HY 248/^' 17/ M 3) ... sent Kwoh (Tso's son) Shing to inform Tsin of the troubles. (Legge, p.404) Jfr % ft f ' (Tso Chuan HY 278/ || 14/1 £ ) Woo announced to Tsin the defeat (which it had sustained from Tso's). (Legge, p.463) On the contrary, the co-verb x_ }jk. cannot be omitted when it is being used to introduce the 01. That is, a sentence such as * f^j- rfi. ^ ^ f % or * ^ ^ It f I is ungrammatical if lu is interpreted as a nominal term. Only a word which has a verbal sense can take the place of i JJA , i.e. * i( $ ffi f H"OR * ffl $a $ If f t • 10. Another alternative seems to be to take the construction ' >//. X' as the first verbal element of the construction 'verbal expressions in series'. Cf. ^ $7 'take pen write characters > write characters with a pen'(Chao 1968:326). Or to take it as an adverbial 333 -construction as in the Japanese structure '* ^ jjj)^ ~Q MA *" ~J It W» "u 1 (Kyo ni wairo o motte Kyochu o motomu) • But it should be noted that in classical Chinese, the construction ' MA X' can be placed after the OB, e.g. jf, |"| & % 4I9 1 fa w^ §i CTso Chuan HY 60/^jf. 12/5-fo, ) '... also requested Nan-king Wan from Ch'in, offering a bribe at the same time' (Legge, p.89). The post-positioning of this structure, as Takashima suggested during a private consultation, may indicate a shift of focus. The above description is virtually applicable to the verbs kao ^ and yu ifp too. In the case of kao ^ , inscriptions having the pattern ' -Jt + (0G) + ( -f ) + OB, VR + OV are not so numerous as those with ch'iu $\, and yji -p^ . In the Sorui, Shima lists some inscriptions which appear to be exceptions to our generalization: a) ?! f Hi £ £ H f« yu / crack / Pin / test / announce / Pi / receive U a-s <?f * ftf (--.I'll) order / to / Ting / two / penned-sheep / split-sacrifice one / ox Ts'ui -f=f 533 e . (Sorui 121.1) V + OG + ^ + OB + OV' , Vg. + OV This inscription seems to go against the rule that wherever there is a specific realizing 'Vg + OV' structure, i. e., xf>|)||> "f" , an OV does not appear right after the structure 'V + (OG) + ( ^ ) + OB'. However, upon checking the rubbing, it is found that Shima's transcription is not totally beyond doubt. See the xerox copy: The upper portion of this piece is mutilated (or unrecoverable for the time being), we cannot ascertain that there was not originally a graph (a specific realizing verb) engraved above the graph ^ , i.e., * % + # + * + ' I+ 5 f " T • cf- this inscription is also listed on p.406.4 of the Sorui. On this page, the graph ^ is transcribed as S , pointing to Shima's uncertainty in reconstructing this partially obliterated inscription. (2) n<p % f ^ |« 1 <p f * ping-wu(?) / crack / next / chia(?)-yin yu-sacrifice / Pi / exorcise ft %*f Pi / exorcise 'ai Chia(?) / Ch'iang / hundrea iang / dismember-sacrifl Ch'iang / dismember-sacrifice / ten penned-sheep Ts'ui %^ 190 This is a very obscure inscription, the graph after f7^ ( l^D ) shows a shape somewhat like p_u (•>- Cto crack); however, fjcj K simply does not make sense. (Shima reproduces the graph [v as j (wu -ip ).) Also, the graph following H. ( ) appears to be the word ^ (to crack), but Q'TJ \j ^ does not make sense either. (Shima reproduces the graph as -j* (chia £p ).) The graph following ^ is also strange, it shows the shape of ( T ten), but + ^ ig is a very strange - 335 -structure (ten and hundred Ch'iang-tribesmen(?)). One may, as implied in Shima's reproduction, supply a horizontal stroke to the graphs -J. and | , that is, ^ -j- ^ ( , the next chia-yin day) and ^ + C T'ai Chia). Not only are there these presumed scribal errors, but positioning of the graphs is also very irregular. See the following enlarged copy where the order of the words is indicated by numerals. It is very doubtful whether this inscription which involves so many errors and irregularities really reflects the inherent structure of the bone language. It is hardly justifiable to take this inscription as a counter-example of the generalization proposed above. Cf. On page 56.2 of the Sorui, Shima reproduces this inscription as • n i v % t J $ t 'I -...tf f 0 f * I f • The dots supplied between the graphs . and X show that Shima too had reservations about the completeness (and presumably the correctness) of this inscription. - 336 -In the cases of liao ^ , yu and yu ty , both the structures 'V + ( ^ ) + OB + OV' and 'V + OV + ( ^ ) + OB' occur, though the former is predominant. These structures coincide remarkably with the syntax of the bronze inscriptions. According to Shen Ch'un-huei's ~'%> studies (1936:399 & 379), 53.3% of the double object sentences in the bronzes take the pattern 'V + OB + DO' while only 5.4% take the pattern 'V + DO + ( ^ ) + OB' (where 'DO' stands for any direct object such as chin , comparable to the OV in the O.B.I. The statistics are provided by the author.) In the case of O.B.I., the less common pattern 'V + OV + ( -j" ) + OB' appears to be a 'marked' form to emphasize the OV. When the number or type of victims is the focal point of a divination, this pattern is more likely to be employed. For example, test / yu-offer / to / Ancestral Mother Chi captive / ^ >$$:(?) We should yu-offer to Ancestral Mother Chi a captive $ (?). Ping 330 (7) should not / yu-offer / captive / to Ancestral Mother Chi Ping 330 (8) We should not yu-offer a captive to Ancestral Mother Chi. < yu-offer / female slave / to ^Ancestral Mother Chi Ping 330 (9) We should yu-offer a female slave to Ancestral Mother Chi. Presumably, Ping 330 (7) is the divination performed the earliest since it is relatively more complete, i.e., the word chen jj9 is kept, while in (8) and (9), it is omitted. By this divination, the Shang want to - 337 -determine what they should do in sacrificing to Ancestral Mother Chi, and the common pattern 'V + -f + OB + OV' is used. When they divine about what particular victim, i.e., a captive or a female slave, should be used, the focal point is shifted and, the OVs are moved to a position in front of the OB. Another example, 1 i~ $ ML H 5 ^ * crack / Ch'Ueh / test / yu-offer j en-hsU to / Ancestor ... We should yu-offer to Ancestor (I) Ping 120 (5) yu-offer / to / Ancestor I / five m •% penned-sheep Ping 120 (6) We should yu-offer to Ancestor I five penned-sheep. three / penned-sheep (It is) three penned-sheep that Ping 120 (7) Ping 120 (8) - - 8 it u yu-offer / one / ox / Ancestor I 'We should yu-offer one ox to Ancestor I. The inscription Ping 120 (7) where all other elements of the sentence are omitted, shows that it is the number (and/or the type) of victims that is the focal point of the divination. Here, we find the pattern 'V + 0V + OB' in Ping 120 C8) . (See also Ping 531 (7), (.9), (10) and Ping 203 (8), (9), (11), (10).) But it should be noted that there are also cases where the focal point lies on the OV and yet the common pattern 'V + ( )+0B+0V' is used, for example,