Introduction
In the context of early Chinese bamboo and wood manuscripts, codicology still is rarely conceived as a distinct field of study. However, the earliest studies that touch upon codicological aspects were published more than a century ago. It seems justified to say that Wang Guowei 王國維 established this field with his seminal Jiandu jianshu kao 簡牘檢署考 (“Study of slips, tablets, address plates, and their inscriptions”). Even if he was not the first scholar to discuss aspects of codicology, his was the first work that not only included evidence from newly excavated bamboo and wood manuscripts but also contained a rather comprehensive discussion of formats of writing support, layout features, and so forth.Footnote 1 Today, most introductory works on bamboo and/or wood manuscripts include a section entitled jiandu zhidu 簡牘制度, “the system of slips and tablets,” a term often used to refer to both textual as well as non-textual aspects of bamboo and wood manuscripts in general.Footnote 2 A description of this “system” therefore also includes codicological features such as “raw materials” (sucai 素材) or “outer appearance/material design/form” (xingzhi 形制 or xingshi 形式).
Early research on bamboo and wood manuscripts was especially deeply rooted in the philological tradition of late imperial China, and certain manuscript-related terms that are well known from received literature were prominent points of discussion. Most of these terms—for example, jian 簡 and du 牘—referred to distinct shapes or formats of pieces of wood or bamboo used as writing support.Footnote 3 Therefore, from today's point of view, they provide valuable evidence for the codicology of bamboo and wood manuscripts. Although the material and textual basis for research has grown significantly over the past century due to the excavation of numerous bamboo and wood manuscripts from pre-imperial and early imperial China, the interpretation of said terms still heavily relies on a very few frequently cited loci classici in the received literature.Footnote 4 And although the most recent works draw on manuscript evidence more extensively, the exact meanings of many terms and the distinctions between them remain difficult to grasp.Footnote 5 Furthermore, diachronic or regional differences in the use patterns of these terms are usually neglected in favor of unifying descriptions.
Another aspect adding to the confusion is that some ancient terms are frequently employed as descriptive codicological terms in modern research. An illustrative example is the word du, which is often found in manuscript publications or excavation reports. It is commonly used to refer to comparatively wide pieces of wood or bamboo, usually with more than one column of writing, in opposition to the more frequently found narrow pieces with only one or at the most two columns of writing (usually referred to as jian).Footnote 6 Their typical English equivalents “tablet” and “slip” or “strip” have already appeared in the text above.Footnote 7 The criteria just named to distinguish between the two, however, merely reflect my own general impression of how the two terms are usually used in Chinese publications. Probably based on an implied consensus, most publications do not define them.
Whether a person living in Qin 秦 or Han 漢 China (3rd c. b.c.e. to 3rd c. c.e.) would have recognized jian and du as something resembling their modern referents is in fact an intricate question. What makes—and more importantly made back then—a du a du (or a jian a jian)? Was it merely the format of the writing support and the number of columns written on it, or was it rather the fact that several jian were usually connected with binding strings to create a manuscript, whereas a single du often constituted a manuscript by itself?Footnote 8 At least for the modern usage of the words, both criteria appear to be somewhat important, but the problem is that they are not perfectly congruent. There are in fact examples of wider pieces with several columns of writing that were connected to other similarly wide or narrower pieces via binding strings.Footnote 9 What were these called at the time of their production? To further investigate the usage of such terms in ancient China could help to arrive at a clearer picture of how and why different terms for writing support developed and according to what criteria certain shapes and formats were distinguished.
Towards this aim, the present article analyzes the usage of two terms that may be described as a complementary pair: the aforementioned term du and the term die 牒. In contrast to what the modern expression jianduxue 簡牘學, or “the study of slips and tablets,” might suggest, in the Qin and Han periods die was among the most commonly used terms for pieces of bamboo or wood that today are usually called jian.Footnote 10 Investigating differences in form and function that can be gathered from the way the terms are employed in both administrative documents and legal prescriptions of the Qin and Han period, this paper argues that du and die were connected to two conceptually different types of manuscripts, namely single- and multi-piece.Footnote 11 This goes back at least to a time shortly after the Qin unification in 221 b.c.e., when the term du may have been newly introduced in order to mirror this distinction terminologically. Although die referred predominantly to pieces of writing support that formed part of multi-piece manuscripts, at least during the Qin period it could likewise refer to pieces that constituted single-piece manuscripts—under certain circumstances. In the following Han period, the “narrower sense” of die—restricted to the context of multi-piece manuscripts—prevailed, whereas the term du rarely occurs in an administrative context. This probably reflects developments regarding the production, use, and storage of administrative manuscripts in the rapidly evolving manuscript culture of early imperial China.
Qin Terms for Pieces of Writing Support
Before focusing more closely on the terms die and du, it is necessary to situate them in relation to other terms for pieces of writing support used during the late Warring States and early imperial period. First, however, some preliminary remarks shall define the exact type of terms that are to be analyzed, because many different words referred to bamboo and wood manuscripts in one way or another, but not all of them are relevant to the present study. For example, throughout the Qin and Han period, the most general word for “writings” of any kind—a category that at the time comprised almost exclusively bamboo and wood manuscripts, plus a probably much smaller number of silk manuscripts—was shu 書, a noun derived from the verb shu, “to write.”Footnote 12 From Qin and Han administrative records, we know that this word was also used as a generic term for “documents” of various types. Depending on whether a document was sealed (feng 封) or not, it could also be referred to with the classifier feng when counted:
司空曹書一封，丞印，詣零陽。Footnote 13
One sealed document from the bureau of the Controller of Works, with [the seal impression of] the Vice Prefect's seal, to be delivered to Lingyang.
In received literature, we find other terms that could refer to complete bamboo or wood manuscripts, such as juan 卷 “roll” or ce 冊 “mat.”Footnote 14 Both terms describe multi-piece manuscripts, the former implying the usual way these manuscripts were stored, the latter being a more neutral description of their physical form. On the other hand, there were also special terms for certain types of manuscripts that consisted of only one piece of bamboo or wood, like fu 符 or quan 券. Although fu and quan were particularly used as certificates and depended on pairs (or trios) of pieces that fit together (with regard to the position of notches) to fulfil their function, each piece has to be seen as a self-contained, single-piece manuscript in its own right. In addition, there were specialized terms for inscribed pieces of wood or bamboo that were only temporarily attached to “manuscripts proper,” for example, “envelopes” or “address plates” (jian 檢) for administrative documents, which could also be furnished with a receptacle to accommodate sealing clay (feng ni 封泥).Footnote 15
The terms that are the focus of this article represent yet another sub-group of all terms used to refer to bamboo or wood manuscripts or parts thereof. This group can be defined as terms referring to individual pieces of bamboo or wood intended or already used as writing support without implying a certain type of text. At this point, no attempt shall be made to distinguish between individual pieces of bamboo or wood that are identical to a complete manuscript (i.e., single-piece manuscripts) and those that merely constitute one part of a (multi-piece) manuscript.
From Qin manuscripts of the late Warring States and the imperial Qin period, that is, the second half of the third century b.c.e., we know that the Qin used at least four different terms that fall into the category defined above: fang 方, ban 版/板, die 牒, and du 牘. The former three are already known in the manuscripts excavated from Shuihudi 睡虎地 tomb no. 11 in 1975. Fang and ban are mentioned in the following passage of a Qin statute affiliated with the office of the “Controller of Works” (si kong 司空):
令縣及都官取柳及木楘（柔）可用書者，方之以書；毋（無）方者乃用版。Footnote 16
Let the prefectural and the metropolitan offices collect willow [wood], as well as [other] wood that is soft and can be used for writing, and make fang out of it in order to write [on them]; if there are no fang available, use ban instead.
Unfortunately, the text does not give any details about the shape or format of fang and ban. The only information provided is that both seem to have been normally produced from wood, not bamboo. It is also clear that ban were considered less suitable for manuscript production, as they were only to be used if no fang were available. The latter therefore appear to represent a more commonly used type of writing support, at least for manuscripts produced in an administrative context. With regard to the huge amount of administrative wood manuscripts from the Qin period that have meanwhile been excavated from well no. 1 at Liye 里耶, it seems most likely that fang referred to an unknown sub-group of these wood tablets of various formats, which often carry several columns of writing, sometimes on both sides.Footnote 17 In his 1985 translation of the aforementioned statute, Hulsewé rendered fang both verbally as “to make something square” and as the adjective “square.”Footnote 18 However, the fact that virtually all pieces found at Liye and other Qin sites (including Shuihudi tomb no. 4, Longgang 龍崗 tomb no. 6, Haojiaping 郝家坪 tomb no. 50, Yueshan 岳山 tomb no. 36) are clearly rectangular rather than square certainly speaks against that translation.Footnote 19 The cited passage in fact seems to contain the only occurrences of the words fang and ban as terms for writing support in Qin manuscripts. However, the two are occasionally used with this meaning in received literature.
久無事，則聘焉。若有故，則卒聘，束帛加書將命。百名以上書於策，不及百名書於方。Footnote 20
If there has been no official business [with another state] for some time, a courtesy call is made on it. If occasion call for it, then, after [the formal business of] the courtesy call is over, a further communication is conveyed in a letter handed in along with a bundle of silk. [A letter of] one hundred characters or more is written on slips tied together,Footnote 21 less than one hundred characters are written on fang.
In this frequently cited passage from the Yi li 儀禮 chapter on “rites of courtesy calls” (pin li 聘禮),Footnote 22 fang is contrasted with ce 策 as to the amount of text the respective carriers could—or should, according to ritual prescriptions—accommodate. Fang are only used if the number of characters does not exceed 100. At least the generally limited amount of writing that fits onto individual wood tablets such as those found in Liye would suggest that fang in the Yi li passage may refer to the same or a similar type of object constituting a single-piece manuscript. Multi-piece manuscripts (ce in the above passage) made of several slips of bamboo or wood are not subject to the same limitation regarding the amount of writing, as their length can generally be adapted by adding further slips. The following passage from the Lun heng 論衡, which contains the terms ban and du, is instructive as to how wood was processed to produce suitable writing support:
斷木爲槧，㭊之爲板，力加刮削，乃成奏牘。Footnote 23
[If you] cut a tree into wood blocks,Footnote 24 split these to make ban, and diligently apply scraping [to the ban], then they become du [that can be used] for memorials/submissions.
The text describes the production sequence from raw material (qian 槧) to a semi-finished product (ban) to the final product (du). 板 has the same phonophoric as 版 used in the Qin stipulation cited above, and the two characters were both used to write the word ban “board, plank.”Footnote 25 This passage suggests that ban were not regularly used for writing, as does the Qin stipulation from Shuihudi. It appears instead that an additional step of processing—according to Lun heng, a sort of scraping—was normally necessary to yield suitable writing support.Footnote 26 It is unclear whether the scraping involved a significant change in format or referred to a polishing of the surface, or both. However, fang was also used to refer to (not necessarily exactly square) surfaces (e.g., of land or skin).Footnote 27 From this, a verbal meaning like “to produce a surface; to level out, plane” may have been derived. In the context of manuscript production the corresponding noun would have to be translated as “levelled/planed piece of wood.” Since the objects to which fang and du refer can both be seen as further processed wooden boards (ban): Can fang and du be equated? This is what Duan Yucai 段玉裁 (1735–1815) suggests in his commentary on the entry for du in the Shuowen jiezi 說文解字:
牘專謂用於書者。然則周禮之版，禮經之方皆牘也。Footnote 28
Du is a special name for [pieces of wood] used for writing. However, the ban mentioned in the Zhou li and the fang mentioned in the Li jing (i.e., the Yi li)Footnote 29 [in the context of writing] all are [referring to the same thing as] du.
While Duan's comment is possibly correct with regard to the equation of fang and du, the passages from Qin statutes and the Lun heng discussed above contradict it, insofar as both sources suggest a clear distinction between boards (ban) on the one hand and further processed du and fang on the other.
The administrative documents from Liye well no. 1 provide valuable additional evidence, as the largest part of these documents consists of individual pieces of inscribed wood. With regard to the Qin statute quoted above, one would expect them to represent fang or, in a few cases, ban. Curiously, both terms are completely absent from the Liye manuscripts—at least from those published so far.Footnote 30 Instead, however, the terms die and du occur rather frequently. In many of these cases, they probably refer to writing support.Footnote 31 What is most interesting is that in some instances the terms undoubtedly refer to the very piece of wood they are written on. In those cases, a definite connection between a piece of writing support and the term that was used to refer to it can be established.Footnote 32 For example, one side of Liye 8-1566 carries a report by the head of the agricultural office:
丗年六月丁亥朔甲辰，田官守敬敢言之：疏書日食牘北（背）上。敢言之。
In the thirtieth year [of the First Emperor of Qin], on day jiachen of the sixth month with the first day dinghai (i.e., July 26, 217 b.c.e.),Footnote 33 Jing, incumbent [Overseer] of the agricultural office, ventures to report [the following]: We recorded in separate entriesFootnote 34 the [categories and numbers of convicts who receive] daily food rations on the back of this du and [hereby] submit it. End of report.
On the other side of the same piece of wood, we find the actual list that is referred to in the report, with five separate entries:
城旦、鬼薪十八人。
小城旦十人。
舂廿二人。
小舂三人。
隸妾居貲三人。
Wall pounders and firewood gatherers: 18 persons.
Minor wall pounders: 10 persons.
Grain pounders: 22 persons.
Minor grain pounders: three persons.
Female bond servants working off fines: three persons.
From the content of the text on Liye 8-1566, the fact that both sides of the tablet carry writing, and the lack of traces of binding strings, one may conclude that Liye 8-1566 constitutes a complete official document on a single piece of wood.
There are further examples of such single-piece manuscripts including the formulation du bei 牘背: a report on the number of households in Qianling prefecture for the years 28 to 33 (219 to 214 b.c.e.) on Liye 8-487+8-2004, as well as a report on officials and conscripts who are to receive food rations during an official journey to the Commander's headquarters (wei fu 尉府) of Dongting 洞庭 province on Liye 8-1517. On these pieces there are likewise no visible traces of binding strings.
Similar documents provide even more examples of direct reference between a text and the piece of writing support on which it was written. However, instead of du bei, some texts contain the expression die bei 牒背. Liye 8-686+8-973 includes a report on the categories and numbers of convicts assigned to work in the armory of Qianling and the tasks assigned to them (zuo tu bu 作徒簿). On one side of the tablet we find a cover letter for the report; on the other we find the actual content in form of a list. Again, there are no traces of binding strings, which also points towards the possibility that this constitutes a single-piece manuscript. The three fragments that have been joined as Liye 8-1069+8-1434+8-1520 contain a comparable document possibly drafted at the same armory as Liye 8-686+8-973, but at a different time.Footnote 35 Two additional cases, in which die probably refers to the very piece of writing support on which it is written, are Liye 8-651 and 9-2352. Both pieces contain a report by the head of Qiling 啟陵 district. The cover letters on the recto refer to an “ex-officio charge” (he 劾) and a “forensic examination” (zhen 診) on “one die,” respectively. In both cases, these seem to be written directly before the cover letter on the same side of the respective piece. The fact that the persons who sent the two reports were also involved in drawing up the ex-officio charge and the forensic examination, respectively, accords well with the combination of the respective texts on the same piece of writing support.
To better compare the pieces referred to by the terms die or du, the relevant data have been collected in Table 1. Based on an analysis of various documents, including many of the examples also discussed above, Sumiya Tsuneko observed that—in contrast to a widespread assumption—there seems to be no connection between the width of a piece of writing support and its designation as du. Furthermore, the way the terms du as well as die are used in the Liye manuscripts suggests that both could refer to pieces of writing support that were used individually and often carried writing on both sides.Footnote 36
Table 1 Overview of Liye pieces with Self-References in TextFootnote 37
For the term die, we can find an additional piece of evidence in the Shuihudi manuscripts. One of the model documents described in the Feng zhen shi 封診式 (“Models for Sealing and Forensic Examination”) contains the following passage:
爰書：某里公士甲等廿人詣里人士五（伍）丙，皆告曰：「丙有寧毒言，甲等難飲食焉，來告之。」即疏書甲等名事關諜（牒）北（背）。Footnote 38
Protocol: A, holder of the first rank, and others from quarter X, 20 men in all, presented the commoner C from the [same] quarter to the authorities. They all reported: “C has ??? (ning 寧) poisonous words; we object to drinking and eating with him. We have come to report him.” Then we recorded in separate entries the name, status and place of residence/birth of A and the others on the back of this die (i.e., the same piece of writing support on which the protocol was written).Footnote 39
This model text of a protocol—although itself written on a sequence of bamboo slips that were tied together and only carry writing on one side—suggests that a “real” protocol could be written on a wider piece (of wood?). The other side of such a piece was—if not regularly, at least in some cases—used to record the personal details of the reporting persons.Footnote 40 This once again shows that die, like du, could refer to pieces of writing support that were used individually.
However, it is important to note that the term die also frequently occurs in the Liye manuscripts preceded by a number. This serves to specify the length of a submission, probably to enable recipients to judge its completeness upon arrival. It cannot be determined without doubt whether a description such as yu jiao nianyi die 獄校廿一牒 “verifications/checks of criminal cases on 21 die” (Liye 8-164+8-1475) referred to several documents of the same type, each written on one die, or to one document of a certain type written on several die. However, it seems more likely that multi-piece manuscripts are meant.Footnote 41 An analysis of all Liye documents containing the term die sheds further light on the matter. In cases where die probably refers neither to the piece of writing support on which it is written nor to an entirely different document, it usually occurs in a cover letter referring to an attachment that consists of additional die (see Appendix A). This use of die seems to be even more common in administrative documents of the Han period.Footnote 42 Although one could argue that the mentioned die may merely have been submitted together with the piece carrying the cover letter rather than physically bound to it, there are some examples where traces of binding strings are visible on “cover letter pieces.”Footnote 43 This suggests that in at least some cases die referred to additional pieces that—together with the piece carrying the cover letter—formed a multi-piece manuscript. A further hint towards this is that cover letter pieces referring to an attachment mostly measure less than 3 cm in width and carry only one or two columns of writing (see Appendix A).Footnote 44 Comparing this with the multi-piece manuscript Liye 8-755 to 8-759—the pieces of which are up to 2.4 cm wide and carry two columns of writing—it seems possible that many if not all of the cover letter pieces once constituted multi-piece manuscripts together with the attached die.Footnote 45
In the Qin manuscripts from the Yuelu Academy 嶽麓書院 collection there are further instances which can be assumed with some certainty to refer to one multi-piece manuscript consisting of several die. For example, two criminal case records are concluded by letters of recommendation for the officials who solved these particularly difficult cases. Both letters state the number of die for the original case records submitted by the investigating officials:
爲奏九牒，上。Footnote 46
[The Judicial Secretary Yang … ] drew up a submission on nine die and submitted it [to the higher authorities].
今獄史觸、彭沮、衷得微難獄，磔辠（罪）一人。爲奏十六牒，上。Footnote 47
Now the Judicial Secretaries Chu, Peng Ju, and Zhong have solved an obscure and difficult case of one person who committed an offence being punished with quartering. They drew up a submission on 16 die and submitted it [to the higher authorities].
That the indicated number of die does not exactly accord with the number of bamboo slips on which the case records are written is probably due to the fact that the records are merely copies of the original case files, and might have been abbreviated, extended, or otherwise modified.Footnote 48 The actual amounts of six and eighteen slips, respectively, are in fact not too far off the numbers given in the attached letters of recommendation (nine and sixteen, respectively).Footnote 49 One may gather that the term die here likely refers to individual pieces that are part of a multi-piece manuscript. As at least the copies of the original records are written on slips of bamboo, one cannot exclude the possibility that bamboo was also used as writing support for the original records and die may therefore also refer to pieces of bamboo.
Even more persuasive evidence for the use of the term die to refer to pieces of multi-piece manuscripts can be found in the recently published fifth volume of the Yuelu Academy manuscripts. The volume contains a Qin ordinance written on a total of eight bamboo slips.Footnote 50 The text on the very last slip reads as follows:
贖。令七牒。尉郡卒令第乙七十六
redemption fees. Ordinance on seven die. Ordinances distributed to/by the Minister of Trials and the provinces, B76Footnote 51
In this case, die must refer to the individual pieces of a multi-piece manuscript, because the actual text of the ordinance that precedes the note “ordinance on seven die” is written on almost exactly this amount of slips. Only shu 贖, the last word of the last sentence of the ordinance text, was written on the eighth slip, together with the note on the number of die and the title of the ordinance. Among the yet unpublished Qin ordinances there can actually be found a near duplicate of this ordinance, likewise written on a total of eight bamboo slips, which differs only in the framing of the ordinance text.Footnote 52 In that case, two more words of the last sentence were written on the eighth slip. It is therefore likely that both copies of this ordinance faithfully record the original number of bamboo or wood slips on which the ordinance was written when it was approved by the emperor. After approval, the original form may not always have been exactly imitated in the process of distribution, such that copies of the original ordinance might slightly exceed seven slips.
To gather a few preliminary findings on the Qin terms for individual pieces of writing support:
1. The four terms (fang, ban, du, and die) were definitely all used to refer to pieces of wood. Only in the case of die do we have hints that it was also used for pieces of bamboo.
2. All four terms were probably (and du and die were definitely) used to refer to pieces of writing support that were employed individually (as single-piece manuscripts) and therefore could conveniently bear writing on both sides. Only die was likewise used to refer to pieces that were part of multi-piece manuscripts and therefore normally carried writing on only one side.Footnote 53
3. The two observations above suggest a tentative distinction between three “terms for pieces of wood that were used as single-piece manuscripts” (fang, ban, and du) and one “term for pieces of bamboo or wood that were used as single-piece manuscripts or as part of multi-piece manuscripts” (die).
4. Extant pieces of writing support that were referred to by the terms die and du do not enable a clear distinction between the two with regard to form (length, width, or number of columns) apart from the fact that du—unlike die—was seemingly not used for pieces longer than 30 cm (see Table 1).
5. As of now, the only possible distinction between fang, ban, and du is that ban was apparently used to refer to coarser, less processed pieces of wood than fang and du.Footnote 54 The latter two may in fact be synonyms.
As has been seen, it is difficult to clearly distinguish the pieces of writing support referred to by fang, ban, die, and du, because they seem to show a certain overlap with regard to form. Furthermore, the ambiguous use of the term die for pieces used in both single- and multi-piece manuscripts seems to prevent a clear-cut distinction according to function. As will be shown below, the use of die is indeed less ambiguous than it appears to be. But before examining further evidence for the distinction of die and du from a Qin ordinance, some additional aspects hinting at a conceptual distinction between single- and multi-piece manuscripts shall be discussed—namely the different forms in which these manuscripts were kept for storage and transport.
For Storage and Transport: Bundles or Stacks Instead of Rolled-Up Mats
One can assume that the fundamentally different forms of single- and multi-piece manuscripts must also have entailed differences in how such manuscripts were kept for storage and transport. Sumiya has drawn attention to a noteworthy detail in the Shuihudi statute discussed above that mentions the terms fang and ban. The cited passage referring to these two types of wooden pieces is immediately followed by prescriptions specifying the materials with which they are to be tied together:
其縣山之多茾（菅）者，以茾（菅）纏書。Footnote 55
In prefectures where there are many sedges in the hills, use sedge to tie documents together.
Sumiya correctly noted that the use of the word chan 纏 suggests that the previously mentioned fang and ban were not tied together consecutively as multi-piece manuscript, since that way of binding is commonly referred to as bian 編.Footnote 56 Instead, a cord was wrapped around a bundle or stack of fang or ban, such that the individual pieces remained self-contained units.Footnote 57 This observation ties in well with a suggestion by Momiyama Akira 籾山明. Based on an analysis of wood pieces among the Liye materials that bear the term shu 束, he proposed that this term referred to “bundles” (Japanese: taba) of tablets or slips rather than pieces of wood with a ladder- or stair-like shape, as formerly argued by the Liye editors.Footnote 58 The topmost part of most pieces on which the term shu occurs was painted black, a feature usually found on pieces carrying titles or similar brief summaries.Footnote 59 Momiyama suggests that shu could indeed be the term used to refer to bundles or stacks of individual documents that were tied together. The aforementioned pieces with descriptive titles containing the term shu may thus have been attached to such bundles as labels.Footnote 60 One can imagine pragmatic considerations with regard to transport or storage as possible motives for this.
The existence of bundles of wood tablets is indeed attested by the archaeological record. For example, the excavation report on Shuihudi Han tomb no. 77 describes wood and bamboo tablets (referred to as du by the authors of the report) that were stored as “bundles” inside a bamboo basket.Footnote 61 Another example is Songbai 松柏 Han tomb no. 1, in which 63 wood tablets (likewise named du by the authors of the excavation report) were “tied in bundles according to content.”Footnote 62 The Liye manuscripts provide further evidence. According to Zhang Chunlong 張春龍, during excavation of well no. 1 there were only three examples of groups of tied-together pieces remaining intact.Footnote 63 Of these, the group Liye 9-1 to 9-12 has been frequently discussed. Each of the twelve tablets constitutes a single-piece manuscript, and an analysis of mirror-inverted imprints of writing has shown that the tablets must have been stored in stacked form before they ended up in the well.Footnote 64 Momiyama considers this to be an example of a bundle (shu).Footnote 65
A close relationship between the words chan 纏 and shu 束 (at least for the second century c.e.) is in fact suggested by the definition of yue 約 in the Shuowen jiezi:
約：纏束也。Footnote 66
Yue means “to tie together in a bundle.”
It is possible that, already in the Qin period, shu was also used verbally in the context of tying bundles of tablets. A hint towards this is the formulation shu fu 束符 in the Yuelu Academy collection of criminal cases, which has been rendered abstractly as “to be bound by a credential” (with the grammatical subject/logical object being the members of a group of five soldiers) in a recent translation.Footnote 67 In the light of the above analysis of the term shu, the translation “to tie the credentials (fu 符) together in a bundle” seems possible as well. Bundles of similar single-piece manuscripts with a certifying function, namely tallies (quan 券), are mentioned in the Liye manuscripts.Footnote 68
The previous section has already clarified that certain terms (i.e., fang, ban, and du) seem to have been employed only to refer to pieces of writing support used as single-piece manuscripts, but not to pieces used as part of multi-piece manuscripts. As can be seen, this distinction is also reflected in the use of the special terms chan or shu (instead of bian) when referring to several single-piece manuscripts that were put together for the purpose of storage or transport.
A Clearer Picture of die and du: Prescriptions for the Drafting of Administrative Documents in a Qin Ordinance
As argued above, die seems to have been the only one of the four Qin terms under investigation that could refer to both pieces of writing support that were used as single-piece manuscripts and pieces that were part of multi-piece manuscripts. Qin regulations concerning the drafting of administrative documents from the Yuelu Academy manuscript collection serve to further elucidate the relation between die and du. For the most part these regulations imply a clear-cut distinction between the two. However, there is one notable exception. As a basis for discussion, the relevant regulations, which are written on ten consecutive bamboo slips, are first cited and translated below:Footnote 69
諸上對、請、奏者，其事不同者，勿令同編及勿連屬，事別編之。有請，必物一牒，各勶（徹）之，令昜（易）智（知）。其一事112而過百牒者，別之，毋過百牒而爲一編，必皆散。取其急辤（辭），令約具別白，昜（易）智（知）殹。其獄奏殹，各約爲鞫113審，具傅其律令，令各與其當比編而署律令下曰：「以此當某某」，及具署辠人毄（繫）不毄（繫）。雖同編者，必章□114之，令可別報、繠卻殹。
Whenever [several] answersFootnote 70 [to enquiries or decisions], requests, or memorials are submitted to a superior, if these are [concerned with] different official matters, do not let [the respective pieces of writing support] be tied together [in the same multi-piece manuscript]Footnote 71 and do not join [previously independent manuscripts] continuously;Footnote 72 tie them as separate [manuscripts] according to [different] official matters. If there is a request, there must be one die per [requested] itemFootnote 73 and each [requested item/request] is to be thoroughly [described]Footnote 74 in order to make it easy to understand.Footnote 75 In case [a submission concerns] one official matter but exceeds 100 die, divide it up, so that no more than 100 die are tied together in one and the same multi-piece manuscript (bian 編). [The separate parts/manuscripts] must in every case be scattered.Footnote 76 Select a key wordFootnote 77 [to be attached to each separate part/manuscript]; let it be brief but complete, distinct and evident,Footnote 78 in order to make [the submission] easy to understand. As far as memorials regarding criminal cases are concerned, in each case briefly draw up a finding of fact and assure that [the facts] have been firmly established, fully append the relevant statutes and ordinances, let each of them be tied together with the applicable precedents,Footnote 79 make a record below the statutes and ordinances stating ‘This is applicable to XYZ,’ and also fully record whether the offenders are kept under detention or not.Footnote 80 Even if [different parts of a submission concerning the same official matter] are tied together [in a multi-piece manuscript, the manuscript] must be [visually marked/separated(?)]Footnote 81 according to text sections in order to enable separate responsesFootnote 82 [with a decision by the higher authorities] as well as accumulatedFootnote 83 rejections.Footnote 84
用牘者，一牘毋過五行。五行者，牘廣一寸九分寸八；115四行者，牘廣一寸泰半寸；三行者，牘廣一寸半寸。皆謹調讙（護）好浮書之。尺二寸牘，一行毋過廿六字；尺116牘，一行毋過廿二字。書過一章者，章□之；辤（辭）所當止，皆陸之，以別昜（易）智（知）。
If du are used, one du shall not [contain] more than five columns [of writing per side]. [If they carry] five columns, du have a width of one inch and eight ninths of an inch (c. 4.4 cm);Footnote 85 [if they carry] four columns, du have a width of one inch and two thirds of an inch (3.85 cm); [if they carry] three columns, du have a width of one and a half inches (c. 3.5 cm).Footnote 86 In all cases diligently take care [to preserve/produce]Footnote 87 a pleasant appearance [of the characters]Footnote 88 while writing [on a du].Footnote 89 On du [with a length] of one foot and two inches (c. 27.5 cm), one column shall not exceed 26 characters; on du [with a length] of one foot (c. 23 cm), one column shall not exceed 22 characters.Footnote 90 If the [text of the] document [consists of] more than one section, it is to be [visually marked/separated(?)]Footnote 91 according to the sections; wherever there is a caesura in the wording/text, there is in every case to be put a mark(?)Footnote 92 for this, in order to distinguish [separate parts of the text] and make them easy to recognize.Footnote 93
爲故書卻上對而復與卻書及117事俱上者，繠編之。過廿牒，阶其方，江其上而署之曰：「此以右若左若干牒前對、請若前奏。」用疏者，如故。118不從令及牘廣不中過十分寸一，皆貲二甲。119
If, for the purpose of submitting an answer [in response] to the rejection of a former document, [that document] is again submitted together with the letter of rejection and the [new/current] official matter (i.e., the answer), tie [all these documents] together accumulatively; if [the resulting multi-piece manuscript] exceeds 20 die, insertFootnote 94 a “section-piece” [for a subheading],Footnote 95 fill the top [of that piece] with inkFootnote 96 and make a record on it stating: “To the right or left of this [piece] so and so many die with the earlierFootnote 97 answer or request or the earlier memorial.” If the list formFootnote 98 is used [the same rules] as before [apply]. If this ordinance is not followed or if the width of the du deviates [from the norm] by more than one tenth of an inch (c. 2.3 mm), this is in every case fined with two suits of armor.
請：自今以來，諸縣官上對、請書者，牘厚毋下十分寸一，二行牒厚毋下十五分寸一。厚過程者，毋得各過120其厚之半。爲程，牘牒各一，不從令者，貲一甲。
御史上議：御牘尺二寸，官券牒尺六寸。
制曰：更尺一寸牘。121
Request: From now on, whenever government offices submit documents with answers or requests to superiors, the thickness of du shall not fall below one tenth of an inch (c. 2.3 mm); the thickness of two-column die shall not fall below one fifteenth of an inch (c. 1.5 mm).Footnote 99 In case the thickness exceeds the norm, it is not allowed that either of them (i.e., du and die, respectively) exceed [the norm] by [more than] half of the [prescribed] thickness. A standard pieceFootnote 100 is to be created, one each for du and die. Whoever does not follow this ordinance is fined one suit of armor.
The Imperial Secretary submits the following proposal [as an addendum to the requested revision/supplementation of the existing regulations]: Imperial du shall have a length of one foot and two inches (c. 27.5 cm); official tally-die shall have a length of one foot and six inches (c. 37 cm).
The decision [of the emperor] stated: Change [the standard length of imperial] du to one foot and one inch (c. 25 cm).Footnote 101 …Footnote 102
These Qin regulations are important for several reasons. First, they provide many valuable details on the norms to which the Qin government tried to hold the production of administrative documents. Second, from an analysis of these prescriptions—and the requested supplements—one can also gather some hints on the circumstances and/or motives behind their promulgation. For example, the proposal of minimum thicknesses for different pieces of writing support might be due to frequent breakage experienced while using comparatively thin pieces. Originally, economic considerations may have motivated attempts to save raw materials to the disadvantage of the stability of writing support. Third, and most important for the present study, the above regulations shed further light on the terminology the Qin used to describe certain codicological or textual features related to manuscripts, including binding techniques, layout, and punctuation. Although the precise meaning of certain terms is still unclear, their diversity testifies to a sophisticated codicological vocabulary. Moreover, the two terms die and du not only frequently occur but are also clearly distinguished with respect to certain aspects of both form and function. These distinctive features can be summarized as follows:
1. Die carry no more than two columns of writing, while du carry three, four, or five columns of writing.
2. Only die are mentioned in connection with the “binding” of multi-piece manuscripts (bian 編).
3. Die have a minimum thickness of about 1.5 mm, while du have a minimum thickness of about 2.3 mm.
So how can this new evidence from a Qin ordinance be brought together with what has earlier been gathered on the four terms fang, ban, du, and die? To begin with, neither fang nor ban seem to play any role as general terms for pieces of writing support. Although the word fang occurs in the regulations once, in this context it must either refer to a piece with a subheading that is part of a multi-piece manuscript, or to a part of text between two subheadings.Footnote 103 It seems very unlikely that it refers to a piece of writing support that is used individually, as in the Shuihudi statutes and the Yi li.
As far as die and du are concerned, all three above-mentioned features suggest that die was mainly used to refer to comparatively narrow pieces that were tied together in multi-piece manuscripts. Direct evidence for this is the use of die together with bian 編, “to tie together/a manuscript produced by tying multiple pieces together.” The comparatively narrow format—suggested by the application of two or less columns of writingFootnote 104—and the lower minimum thickness also hint towards the use of die as part of multi-piece manuscripts. Thinner pieces with less volume and weight were probably beneficial for the handling and storage of these manuscripts. At the same time, a manuscript consisting of several layers of thin pieces rolled up together would likely be at least as robust as a slightly thicker individual piece used as a single-piece manuscript. The latter are obviously referred to as du in the Qin regulations under discussion. In sum, this means that die and du are distinguishable with respect to both form and function and appear to be connected to two conceptually different types of manuscripts: the multi- and the single-piece manuscript, respectively.
However, the expression guan quan die 官券牒 (“official tally-die”) on slip 121 of the Qin ordinance confirms an observation, already made on the basis of other Qin manuscripts above, which somewhat contradicts this clear-cut distinction between die and du. As has been shown, there is evidence in both the Shuihudi as well as the Liye manuscripts that die was sometimes used to refer to pieces of writing support that were used as single-piece manuscripts. The function of quan, bi- or tripartite tallies with corresponding notches that were used as certificates of transactions, suggests that every part of a quan constituted a single-piece manuscript in its own right. To tie such a document into a multi-piece manuscript would have made it difficult to verify the validity of a transaction by putting together parts kept by different parties. Still, the Qin regulations under discussion suggest that the pieces used as quan were referred to as die rather than du. Footnote 105 The reason for this might be that, as far as their format or more precisely their length to width ratio is concerned, quan are more similar to the narrow die that were used in multi-piece manuscripts than to the wider (and relatively short) du used for single-piece manuscripts.Footnote 106 According to the preface of volume 1 of the Liye manuscripts, the quan excavated at Liye have a length of 37 cm, or “one foot and six inches” (yi chi liu cun 一尺六寸).Footnote 107 A cursory glance at a compilation of all pieces bearing notchesFootnote 108 in conjunction with the respective photographsFootnote 109 shows that there are at least 29 examples of complete—or already completely reconstructed—quan, the length of which can therefore be determined (see Appendix B). As can be seen, all quan are made of comparatively long and narrow pieces of wood. Their length varies between 34.6 and 38.2 cm, while they are only between 1.1 and 2.2 cm wide. Furthermore, none of them bears more than two columns of writing or has writing on both sides. All these features closely resemble pieces commonly used for multi-piece manuscripts. And exactly this may be the reason why these pieces, even if used for a special type of single-piece manuscript, were still referred to as die. In fact, the four pieces with the self-reference die discussed above (see Table 1) are comparatively long and narrow too, measuring more than 15.1 × 1.7 cm (Liye 8-651), more than 32.3 × 1.9 cm (Liye 8-686+8-973), 46 × 3 cm (Liye 8-1069+8-1434+8-1520) and 47.1 × 3.1 cm (Liye 9-2352), respectively. In these cases, not even the fact that they carry writing on both sides in up to four columns seems to have dissuaded the person who drew up the document from referring to the pieces as die. The form criterion (long and comparatively narrow) apparently outweighed the function criterion (single-piece manuscript) in case the two conflicted with regard to the choice of designation.Footnote 110
The Emergence of the Distinction Between die and du in the Qin Period
It is possible that the Qin at some point newly introduced the term du, because it does not seem to occur in either excavated or received sources from before the imperial Qin period. The earliest occurrences can be found in two passages of the Zhanguo ce 戰國策 and the Zhuangzi 莊子.Footnote 111 If du really refers to the same thing as fang, as hypothesized above, then du may have replaced the latter term in specific contexts. Fang certainly continued to be used as a codicological term but apparently referred specifically to pieces of multi-piece manuscripts carrying a subheading or to a part of text delimited by such a subheading, as shown by the Yuelu Academy regulations. The earliest Liye document on which du occurs can be dated to July 26, 217 b.c.e. (Liye 8-1566).Footnote 112 This is—at least according to my own research—the earliest precisely datable occurrence of du in either received or excavated sources. Could the term du have been introduced as one of various terminological changes in the course of the Qin unification in 221 b.c.e.? The most prominent piece of evidence for such lexical changes from among the Liye manuscripts is 8-461, a wooden board measuring roughly 28 × 13 cm, which originally must have described more than 50 changes of certain terms (or character forms).Footnote 113 The board is fragmented, and the writing is not entirely decipherable, but at least in the legible part no mention is made of du. However, there likely were additional lexical changes.
Xing Yitian has observed a peculiarity regarding the use of the terms ban 半 and fa 發 in the Liye documents, which according to context must both mean “to open”Footnote 114—to break the seal of a document. He noted that ban only occurs in documents dated to between the twenty-sixth and the thirty-first year of the First Emperor of Qin (i.e., 221 to 216 b.c.e.), whereas fa only occurs in those of the thirty-first year or later. Accordingly, he hypothesized that there might have been another major lexical reform around the year 216 b.c.e.Footnote 115 Checking Xing's data, I could not find any documents from the thirty-first year in which ban occurs in the sense “to open.” Instead, the latest example seems to be Liye 8-1566 (thirtieth year, July 26, 217 b.c.e.), which happens to be the same tablet that bears the earliest occurrence of du. The earliest example of fa can actually be found on Liye 9-981, which likewise dates to the thirtieth year, more precisely October 20, 217 b.c.e.Footnote 116 Therefore, it seems probable that the standard term to refer to the opening of a document was changed from ban to fa sometime between July and October 217 b.c.e.Footnote 117 The motive behind this change may have been the graphical similarity between the character forms used to write ban 半 “opened by” and shou 手 “handled by,” another expression frequently occurring in the Liye documents.Footnote 118
Du may have been introduced at roughly the same time, because no mention of the term is made in Liye documents dated to before 217 b.c.e. Unfortunately, the Qin ordinance on documents discussed above bears no date. However, the prescriptions regarding the “official tally-die” (guan quan die) contained in them may be used to narrow down the possible date of its enactment. The ordinance text notes that the Imperial Secretary proposed to fix the length of official tally-die at “one foot and six inches” (yi chi liu cun, or about 37 cm) in addition to changes “requested” (qing 請) by an unnamed person—probably a Chancellor (chengxiang 丞相). A glance at the table in Appendix B shows that nearly all examples of quan excavated from Liye match this 37 cm standard. However, a closer look reveals that this is only the case for tallies from the first day of the thirty-first year of the First Emperor (November 4, 217 b.c.e.) onward.Footnote 119 Of these, nearly 80 percent (22 out of 28) have a length between 36.5 and 37.5 cm, while only 6 are slightly longer (37.6 to 38.2 cm). The maximum deviation from the standard of 37 cm is 1.2 cm. There is only one complete tally from before the thirty-first year, which is Liye 8-1551 from the twenty-seventh year (exactly February 5, 220 b.c.e.). In fact, that piece is significantly shorter than all the other tallies. Measuring only 34.6 cm, it corresponds to “one foot and five inches” (yi chi wu cun 一尺五寸, or 34.65 cm) rather than the “one foot and six inches” stipulated in the Yuelu Academy ordinance.Footnote 120 Unfortunately, the only other two tallies predating the thirty-first year, viz. Liye 8-1690 (twenty-ninth year) and Liye 8-1647 (thirtieth year), are only fragments, of which the other parts have not been identified yet. Accordingly, their original length is unclear. However, for the time being one might tentatively date the enactment of the supplemented version of the above Yuelu Academy ordinance to the period between February 220 b.c.e. and November 217 b.c.e. As the original regulations that were to be supplemented already contain the term du, that term must have been introduced at some point before. Whether that was shortly before the enactment of the supplemented version—at the latest in November 217 b.c.e.—or even before 220 b.c.e. cannot be determined at this point. However, both suggestions seem generally possible, as the former would roughly coincide with the lexical change from ban to fa, while the latter would link the introduction of du to other lexical changes directly following the Qin unification in 221 b.c.e.
It seems that the Qin distinguished between du and die at the latest by 217 b.c.e. Pieces of writing support referred to as du were usually used for single-piece manuscripts and often carried writing on both sides in three or more columns. Pieces referred to as die were normally used for multi-piece manuscripts and often carried writing on only one side in no more than two columns. As has been shown, pieces used for single-piece manuscripts were occasionally referred to as die on the basis of their form. However, form was not necessarily the primary criterion for the choice of designation. The different ways in which the two types of manuscripts were kept for storage and transport suggest that function (single- vs. multi-piece manuscript) figured prominently as well. Possible etymological relations of the terms die and du with other words provide further clues on this. For example, Schuessler suggested a relation between die 牒 and die 褶, “double, lined garment,” as well as die 疊, “double, accumulate.”Footnote 121 This would fit the previous observation that die 牒 are normally “accumulated” (i.e., tied together) in multi-piece manuscripts. With this in mind one can hypothesize that du 牘 *lˤokFootnote 122 may be a cognate of du 獨 *[d]ˤok “alone; self-contained, independent.”Footnote 123 If this is actually the case, then the basic meanings of die and du as codicological terms could be “pieces tied together in multi-piece manuscripts” (or other pieces closely resembling them) and “pieces used as single-piece manuscripts,” respectively. For pragmatic reasons the translations “slip” or “strip” for die and “tablet” for du may be more suitable, but care should be taken not to confuse this terminological use with non-terminological uses in which the English terms are not necessarily the exact equivalents of die and du.
It is remarkable that—at least based on the cited Qin regulations—there does not seem to have been a preference for either multi- or single-piece manuscripts in the production of official documents.Footnote 124 Although multi-piece manuscripts were of course widely used during the Qin period, as the many bamboo manuscripts from tombs show, single-piece manuscripts seem to have been very common in administrative practice—at least if the Liye manuscripts published so far can be considered representative.
Further Developments in the Han Period and Possible Reasons
In received literature from the Han period, both die and du are frequently mentioned. Occasionally, the two occur together, which might suggest that the terms continued to be used to refer to single- vs. multi-piece manuscripts as two basic types.Footnote 125 However, whereas the term du appears in Han received literature, it is all but absent from actual administrative documents of the time.Footnote 126 This presents a stark contrast to the term die, which occurs more than 250 times in the documents from Juyan 居延 and Jianshui Jinguan 肩水金關 alone. In more than 70 percent of these cases die appears as part of the phrase ru die 如牒, literally “as [written] on the die.” It has been shown that die in these cases means the pieces of writing support that are―or rather were, originally―attached to the ones on which the text including ru die is written: cover letters. Furthermore, die could refer to pieces carrying either one or two columns of writing. These were named zha 札 and liang hang 兩行, respectively, before they became part of a multi-piece manuscript.Footnote 127 Therefore, a suitable English translation for the expression ru die would be “as [written] on the [attached] die” or simply “as in the attachment.”Footnote 128 Similar cases where attached pieces are referred to as die in a cover letter can also be found in the Qin documents from Liye (see Appendix A).
The difference in usage of the term die in the Qin vs. the Han period seems to be that, as has been demonstrated, form still had an influence on the choice of designation as du or die in the Qin period, whereas the distinction seems to have evolved into a clear-cut functional one in the following Han period.Footnote 129 This means that die occasionally referred to pieces used as single-piece manuscripts in the former, whereas it is exclusively used to refer to the individual pieces of multi-piece manuscripts in the latter. Sumiya pointed out that, although the second usage of the term die is indeed suggested by the Juyan manuscripts, a passage in the Han “Statutes and Ordinances of the Second Year” (Ernian lüling 二年律令) from Zhangjiashan 張家山 tomb no. 247 shows strong parallels to the Qin usage of die. The passage, from the “Statutes on Agriculture” (Tianlü 田律), reads:
官各以二尺牒疏書一歲馬、牛它物用稾數，餘見芻稾數，上內史，恒會八月望。Footnote 130
Each office is to write down in separate entries on die of two feet (i.e., around 46 cm) [in length] the amount of straw used in one year by its horses, cattle, and other animals, [as well as] the remaining amount of hay and straw, and submit it to the Scribe of the Capital Area. The deadline [for submissions] is always the full moon of the eighth month.
A “die of two feet in length” with a report “in separate entries” would in fact precisely describe the Liye piece 8-1069+8-1434+8-1520 discussed above, which actually constitutes a single-piece manuscript.Footnote 131 Because in Han administrative documents die usually refers to narrow pieces with only one or two columns that are tied together as multi-piece manuscripts, Sumiya suggests that this Han stipulation might be based on earlier Qin regulations.Footnote 132 This definitely seems possible in light of the substantial textual similarity between certain Qin and Han statutes.Footnote 133 Another possibility is that the obvious changes with regard to the use of the term die did not occur at the beginning of the Han period, but instead postdate the time of the Ernian lüling—generally dated to 186 b.c.e.Footnote 134
Although single-piece manuscripts continued to be used in administrative practice throughout the Han (206 b.c.e–220 c.e.) and into the Three Kingdoms period (220–80 c.e.), multi-piece manuscripts seem to have figured far more prominently. Finds of administrative documents from these periods that were all excavated in the same area in present-day downtown Changsha (Hunan province) provide a suitable set of data for comparison. Sorted according to period they can be summarized as follows: (1) Zoumalou 走馬樓 well no. 8: The manuscripts were roughly dated to the early reign of Emperor Wu of Han 漢武帝 (r. 141–87 b.c.e.). While the excavation report of 2005 mentions only pieces with one or two columns of writing, Song Shaohua 宋少華 describes four basic types. Three of these have one or two columns of writing, and one has three columns. Still this suggests that the former three types of pieces, which were likely once part of multi-piece manuscripts, represent the majority, as the latter type was not even mentioned in the original excavation report.Footnote 135 (2) Wuyi guangchang pit no. 1: The manuscripts date to around 100 c.e. and mostly consist of pieces with two columns, probably once part of multi-piece manuscripts. A few examples of single-piece manuscripts were also found, e.g., CWJ1③:291, CWJ1③:169. The exact ratio of single- vs. multi-piece manuscripts is as yet unclear, as only a selection of the manuscripts has been published to date.Footnote 136 (3) Dongpailou 東牌樓 well no. 7: The manuscripts date to the reign of Emperor Ling of Han 漢靈帝 (r. 168–89 c.e.). Among the overall 205 inscribed pieces there are various single-piece manuscripts. However, these seem to be mostly private letters or official letters with an informal nature.Footnote 137 (4) Zoumalou well no. 22: The manuscripts date to 220–37 c.e. and most pieces were originally part of multi-piece manuscripts. A few examples of single-piece manuscripts can be found, e.g., J22-2695.Footnote 138 Based on an analysis of documents excavated from the northwestern border regions of the Han Empire, Sumiya has pointed out that the use of single-piece manuscripts, which she refers to as tandoku shiy ō 単独使用, was largely restricted to private letters or official letters with an informal nature, or the so-called “notes” (ji 記).Footnote 139 Taking the “register/account” (bu 簿) as an example, she furthermore demonstrated that this type of document was often drawn up as single-piece manuscript during the Qin period, whereas in the following Han period it usually constituted a multi-piece manuscript.Footnote 140 Takamura Takeyuki 高村武幸 delivered additional evidence for this tendency and argued that the reign of Emperor Wu of Han may be seen as a transitional period during which the use of single-piece manuscripts became largely restricted to a few special circumstances. For most types of documents, single-piece manuscripts seem to have been replaced by multi-piece manuscripts. This is why the former do not figure prominently among the Juyan and Dunhuang manuscripts, most of which can be dated to a later period.Footnote 141 This would also explain the absence of the term du in administrative documents.
This demise of single-piece manuscripts from the Qin period onwards—at least in an administrative context—inevitably raises the question of the reason for this development.Footnote 142 Scholars have advanced several hypotheses. For Sumiya, the main reasons behind the choice of certain types of writing support—with regard to both form and material—are regulations such as those found in the statutes excavated from Shuihudi tomb no. 11 and Zhangjiashan tomb no. 247. At least for the Qin period, these seem to suggest that wood was normally used to produce writing support for administrative documents and that single-piece manuscripts were the most common type. The Liye documents in fact seem to confirm this.Footnote 143 Sumiya further argues that this preference might result from considerations regarding the integrity of administrative documents. Single-piece manuscripts do not come with the danger of binding strings dissolving and parts of a document getting lost, which may be seen as a major advantage of single- over multi-piece manuscripts.Footnote 144 However, considering the Qin ordinance discussed above, there does not seem to have been a prescribed preference for either single- or multi-piece manuscripts. Takamura argues that the degree to which single- vs. multi-piece manuscripts were used in the northwest of the Han Empire (Juyan and Dunhuang) may not have institutional causes. He suggests that the availability of raw materials, individual or group habits, different production techniques, and/or the amount of text to be recorded probably had a strong influence on the choice of writing support (and accordingly the type of manuscript).Footnote 145 Sumiya's hypothesis is apparently the only explanation yet proposed as to why the Qin administration seems to have made use of single-piece manuscripts so frequently. The possible reasons for the later preference of multi-piece manuscripts, however, are numerous:
First, multi-piece manuscripts are structurally more flexible. They enable the separation of one manuscript into several smaller units. At the same time, they facilitate compiling formerly separate manuscripts and also allow expansion if need be.Footnote 146 Furthermore, it is possible to replace individual pieces of multi-piece manuscripts if, for example, extensive corrections or changes are necessary. All this would be impossible or at least impractical with single-piece manuscripts—usually consisting of a wide piece of writing support with writing on both sides—and would probably have made an additional copying process necessary.Footnote 147
Second, multi-piece manuscripts require smaller pieces of writing support. Although it is theoretically possible that a whole tree trunk would have been used solely to produce writing support, it seems more likely that the most valuable parts of a tree were used for other purposes, first and foremost the construction of buildings. Therefore, writing support was likely produced from the remaining pieces, which would be considerably smaller. Pieces of writing support designed for only one or two columns of writing could even be produced from very thin raw materials the size of a branch—with few other possible uses apart from as firewood. This would be especially advantageous in areas with very limited supplies of raw materials.Footnote 148 Wider pieces would require raw material of a size that might also be suitable as timber. In addition, bamboo might even have an inexpensive alternative to wood for the production of writing support suitable for multi-piece manuscripts.Footnote 149
Third, multi-piece manuscripts facilitate mass production. One factor informing this is the smaller format of the required pieces of writing support, which enables a wider selection of raw materials to be used. Another is the lower variety of formats. To produce only two basic types of writing support used for multi-piece manuscripts—zha and liang hang—is probably easier and therefore economically much more favorable than to produce six additional types of wider pieces, that is, du used as single-piece manuscripts with three different widths and two different lengths.Footnote 150 Of course, zha and liang hang used in administrative documents were also produced in different lengths, but by far the most common length was one Han foot (or roughly 23 cm).Footnote 151 The lower variety in width must have greatly facilitated the production of standardized writing support.
Concluding Remarks
An analysis of the usage of the terms die and du in excavated and received sources from early imperial China has shown that the two terms are connected to conceptually different types of manuscripts, namely single- and multi-piece manuscripts, which also entail differences in the way they are kept for storage and transport. This functional distinction may be reflected in etymological relations of the two words with other words of the semantic fields “duplication, accumulation” (die) and “alone; self-contained, independent” (du), respectively. Furthermore, there apparently was a distinction with regard to form, which seems to have outweighed the functional distinction in certain conflicting cases—for example, when comparatively long and narrow pieces were used as single-piece manuscripts (e.g., tallies referred to as die). It is possible that this partial overlap of the meanings of die and du in Qin usage resulted from the introduction of du as a novel term sometime after the Qin unification. In the following Han period, multi-piece manuscripts prevailed over single-piece manuscripts. This is also suggested by the frequent occurrence of the term die, which, at least by the time of Emperor Wu, referred exclusively to pieces of writing support that were part of multi-piece manuscripts. It has been argued that there are numerous pragmatic and economic reasons that could explain the preference of multi- over single-piece manuscripts and that would moreover fit well into the setting of a gradually consolidating empire with an ever-growing volume of bureaucratic record keeping.
Considering the economic value of the pieces of writing support from which the two types of manuscripts were produced, the comparatively wide pieces usually used for single-piece manuscripts must have been more expensive than the narrower pieces used for multi-piece manuscripts.Footnote 152 Exactly this may be one of the reasons why single-piece manuscripts did not vanish entirely. It probably is no coincidence that they continued to be used in contexts where courtesy and respect must have been much more important than the pragmatic and economic considerations governing the production of administrative documents: the exchange of private (or at least informal) letters and greeting tablets.Footnote 153
Appendix A: Overview of pieces from Liye Well no. 1 mentioning die Footnote 154
Appendix B: Overview of complete quan from Liye Well no. 1Footnote 155
