A NEW COMPREHENSIVE MONOGRAPH ON KHITAN

This three-volume book is the latest comprehensive monograph on the Khitan monuments written in the Khitan Small (or Assembled) Script. The last three similar works written in the 21st century with a similar aim are the works of Chinggeltei (2002), Wu Yingzhe (2012) and Liu Pujiang and Kang Peng (2014). The work under review (henceforth CWJ) is in all respects a richer and more comprehensive work than the former ones. It is the publication of the latest results of the work of a research group founded and led for a long time by the famous professor Chinggeltei (†27 December 2013). His work was continued by his pupils and two of them are the co-editors of this work. The three volumes of the book are titled as follows: Vol. 1: Further Research on Khitan Small Script and Photocopies of the Monuments [English translations by the reviewers] (Qidan xiaozi zai yanjiu ji ziliao tuban 契丹小字再研究及资料图版), Vol. 2: A Collection of Khitan Small Script Corpus (Qidan xiaozi ziliao zongbian 契丹小字资料总编), Vol. 3: Indexes (Suoyin 索引).

of the history of their research. Glyphs are paired either with their reconstructed phonetic manifestations (ni yin 拟音) or with their hypothetical readings (jiadingxing niyin 假定性拟音); e.g. No. 69  has a reconstructed reading: ri; No. 68  has hypothetical readings: biʧ or us (p. 178). Glyphs are marked by cardinal numbers. Nos. 9 and 10 are exceptions, since their readings are under No. 1. A reference to such dispositions would have been adequate. Features like vowel harmony, vowel reading rules, traces of Chinese -p, -t, -k codas of the entering tone as represented in KSS are thoroughly discussed in Part 1 (pp. 128-148).
Chapter 4 (第四章) on pp. 354-600 (and not starting on p. 355 as it is written in the English Table of Contents) deals with the semantics of KSS monuments. The authors analyse the semantic features of KSS materials as combinations of single glyphs. They distinguish three types of morphemes: (1) units having their own meaning; (2) units that have no meaning when isolated, while they have meaning in compounds; these may have either lexical or grammatical functions; (3) sequences with both lexical and grammatical functions. The authors separate two types of semantic groups of the glyphs based on their functionality, i.e. ones carrying 'primary meaning' (zhuziyi 主字义) and ones having only secondary (or 'attached', cf. the English translation of the Table of Contents 'Additive Elements' fuziyi 附字义) meaning' (or rather 'function'), a new terminus technicus introduced by the authors for glyphs representing suffixes or particles. The former category is described in Part 1, pp. 355-488, the latter on pp. 489-594, supplemented by charts of glyphs of the second category already deciphered on pp. 595-600. Examples for the above partition: glyph representing a lexeme:  DAY (p. 403); glyph representing a grammatical function: the plural marker  li (pp. 538-539).  DAY was earlier treated as a logograph, a glyph only the meaning of which is known. In this case, we learn from the Liaoshi 遼史 that the word was pronounced as /närä/. The claim that -li is a plural marker is interesting, but needs further corroboration.
Photocopies of (the rubbings of) the extant KSS monuments constitute an unnumbered Chapter on pp. 601-864 (and not starting on p. 603 as it is written in the English Table of Contents). The inclusion of good quality photocopies in most cases is of great help to scholars having no access to the monuments; however, as some inscriptions are simply too large to fit the page in a reasonably legible size, the help is not always effective (e.g. p. 659 or p. 834).

Volume 2
The second volume contains the corpus of the known KSS monuments presented in four groups according to the time of their first publication (1: 1922-1977, pp. 866-1022; 2: 1977-2002, pp. 1023-1223; 3: 2002-2010, pp. 1224-1448; 4: after 2010, pp. 1449-1592) saving a fifth chapter for the minor texts of metal (bronze mirrors, coins) and stone tablet inscriptions (pp. 1593-1615). The discussion of the material is structured around the following pattern: 1. A short introduction of the monument expounding the historical background, place and date of unearthing, wherever these are identifiable, as well as physical and textual description. A great change compared to earlier mainland Chinese works is the practice that the old Chinese parts of the Khitan-Chinese bilingual texts are not rendered in simplified Chinese characters in the transcripts, giving thus no ground for misinterpretations. The glossing of the Khitan texts is in simplified script as pointed out above, but that does not restrict its usability. The transcript of Khitan texts in Menksoft Khitan fonts provides great readability.
The Table of Contents along with the Preface and the Postscript translated into English at the end of Vol. 3 will be useful for scholars not reading Chinese; however, for such a generously edited opus magnum, a basic proofreading of these few pages would have been welcome.
The volumes are printed in the traditional Chinese way, i.e. the book reads from right-to-left. A somewhat confusing method of pagination is applied throughout the volumes. Solely Chinese numerals are used, but parts not belonging to the main text, such as the Preface, Tables of Contents, Postscript, restart pagination from 1, whereas the pagination of the main text runs through the volumes. The whole trend is then broken at the end of Vol. 3 where the pages containing the English translations are numbered the opposite way around in a Western, left-to-right manner, starting from the very last page as 1 until p. 19, which is then preceded (followed?) by p. 2336. Thus Vol. 3 has two sets of pp. 1-9, and there is no p. 1616 marked either at the end of Vol. 2 or at the beginning of Vol. 3.
The work has two great merits: it made accessible more material in KSS than any earlier publication and it is an outstanding help for those who work with the Khitan material.
In the last comprehensive overview (Apatóczky and Kempf 2017: 115-122), 41 major inscriptions were only mentioned. In this book, however, the 44 major and 17 minor inscriptions are also published. In this case, 'publication' means that we find photocopies, transcriptions in KSS fonts, and the transcriptions are in many cases glossed by Chinese translations. The whole corpus was Romanised, the Romanisation of the inscriptions has been digitalised. The corpus includes more than 100,000 items. In the Index, the units are listed according to their current number in the List in growing order. The Index contains 10,407 different units. A unit may be a basic word or a derivation, or a suffixed form or a morpheme with only grammatical function. As it is known, the KSS is written in 'boxes', or are 'assembled', hence the other name of the KSS: Assembled Script, a term first used by G. Kara (1987). A unit may consist of maximum 8 glyphs in the following order: a b c d e f g h For technical reasons, in the case of Romanisation a linear transposition has hitherto been used with dots denoting that the glyphs of the item pertain to the same 'box', e.g. the example above would be: <a.b.c.d.e.f.g.h>. Now in CWJ a hyphen is inserted <a-b-c-d-e-f-g-h>. The items also have a current number. The GREA T  (373) xomur was a type of coffin, <dor>  (218) is a logograph for SEAL, used in the meaning 'ceremonial', that is 'the ceremonial coffin of the tomb'.
In the Index, all occurrences are given, and thus the frequency of a unit in the corpus can be easily identified. In the Index, one can find many hapaxes, units that occur only once; and, on the other hand, units that occur several hundred times. Frequency will be an important aid in deciphering the language; however, one has to distinguish Khitan words and syllables used in transcriptions for Chinese. The deciphering of the script is also advancing. The last overview of the stage of deciphering the KSS can be found in the publication by Wu and Janhunen (2010: 38-48, see also Shimunek 2017: 419-445). In fact, at present we have to work with the so-called Romanisation of the glyphs or graphs. The glyphs are drawings and are used for rendering, in most cases, syllables. Romanisation is reflecting the view of earlier scholars on the possible reading of the glyphs. The reading, as it is known, was first attached to glyphs which occurred in transcriptions of Chinese proper names and words. Their value depends on the reconstructed pronunciation of the original Chinese. As we see, Romanisation is neither a phonetic, nor a phonological reproduction of the Khitan words. For the time being, we have to accept the opinion of Wu and Janhunen (2010: 39) who stated that 'they [i.e. the Romanised transcription of the glyphs] are best understood as intermediate labels for the Khitan characters concerned'.
In comparison to Wu and Janhunen's work, there are some small changes concerning the consonantal Latin stock used. The Romanisation <ts> has been changed to <ch>:  (031) (only on p. 40 of WJ <tz>, but not in their Index), which has been changed by CWJ to <chi>. The Romanisation <dz> in the case of  (354) has been changed to <s 2 >, however, in the case of  (104), it has been kept as <dz 2 >, while there is no <dz> in the system. These smaller inconsistencies do not hamper the work.
One would expect that with the inclusion of the new inscriptions many new glyphs appear. On the contrary, WJ had in their List of Glyphs 459 items, while the last glyph in the List of CWJ is numbered 378. True, there are some new items inserted. CWJ introduced a new custom. Until this publication, if a new glyph was found, it was put at the end of the List and got a new number, which was a safe, even if a little cumbersome, procedure. According to the new system, when a newly found glyph appeared, it got an extended number attached to an already existing one.
In most cases it denotes an alloglyph as in:  (261.1) which may be an alloglyph of  (260) <l>. While (261) is one of the most frequent glyphs, (261.1) occurs, at least as initial, only once in Yong30-11, where the facsimile is not quite clear. Another example is  (361.1), which is an alloglyph of  (072), and occurs also only once in Hui2-2. In such cases, it seems to be reasonable to give an extended numeral to a new, uncertain item. This is, however, not always the case.
It was among others Wu who supposed that the so-called dotted variants, or at least many of them, signalised grammatical gender; so far the dotted variant has marked the male and the undotted one the female gender (of course, the absence of a dot may be due to many other reasons, e.g. simple negligence). From the cases of the numerals and the colours we know that in Khitan these gender variants may have different suffixes (see Róna-Tas 2016: 131), but in the known cases the dotted and undotted variants pertain together. It is unlikely that  (082) <üe> and its undotted form  (082.1) would not pertain together. For the latter form, CWJ gives the Romanisation <aju> (aʤa/aʤu). On pp. 186-187, there is a detailed indication for the reading. Though the reasoning of CWJ is not very convincing, of course, we cannot exclude that their suggestion may be valid. If they are right, the extended numeration (082.1) raises more problems than it solves.
In the monograph of WJ (2010: 38), there are 38 items listed as allographs or alloglyphs of other characters. All of them are included in CWJ, too, but in numerous cases an alloglyph has no reading in CWJ as:  (350, occurring four times), which is an alloglyph of the very frequent  (254) <d>, but (350) has no reading, though with the highest probability it has to have the same reading: <d>.
Some items are not even alloglyphs, but in fact identical, like  (171) and  (215). In this case (215) (1101), <es> 10 times, <ul> also 10 times. The earliest occurrence of <es> is in the inscription Guang (1053) where it occurs 16 times. The glyph <ul> (336) is first occurring in Yu (1072), also 16 times. In the whole corpus, the glyph <es>  (146) occurs 465 times, while <ul>  (366) only 130 times. Thus, before hasty conclusions are drawn, further research is needed to clear their respective functions. This case may show that the monograph under review does not only bring new solutions, but it also raises new problems and gives food for further investigations.
As a summary, we may conclude that the work of Chinggeltei, Wu Yingzhe, Jiruhe and their colleagues is a great achievement. It contains almost all of the hitherto known Khitan texts written in the Khitan Small or Assembled Script. The presentation of the material is at the highest level of modern scholarship, and the work will be fundamental for any further research.