敦煌佛经残卷缀合释例

Key Factors of Patching up Fragmentary Dunhuang Buddhist Scriptures

残卷的缀合是敦煌文献整理研究的基础工作之一,缀合时必须注意把握那些对残卷缀合起关键性或决定性作用的因素。残卷缀合有12个关键因素,包括内容相连、碴口相合、字体相同、书风近似、抄手同一、持诵者同一、藏家同一、行款近同、校注相涉、污损类同、版本相同、形制相同等。在具体的缀合实践中,既要紧紧把握制约残卷缀合的关键点,又要综合分析各方面的线索,使残卷的缀合建立在可靠的基础之上。

Most of the Dunhuang Manuscripts are fragments,which brings great difficulty tonaming,dating and further studying them.Among the fragments,some were split apart from thesame document. If these fragments can be patched up, much more information about thedocuments will be found.Therefore,patching-up is part of the groundwork of philological studyon Dunhuang Documents.There are several key factors to take into consideration in patching upthe fragments.In this paper,12 key factors will be introduced with examples from Buddhistscriptures.1.Successive content.If the contents in different fragments are connected or related to eachother,these fragments may have been split from the same document.For example,the last wordof each column in BD 1 1242 is followed by the first word of the next column in P.5587（1 1）.Thisinformation plays a key role in rejoining BD 1 1242 and P.5587（1 1）.2.Matched edges. The edges of different fragments have different shapes, and manyfragments have incomplete characters on the edges.If the edges and incomplete characters indifferent fragments can be matched,the fragments can be patched up.For example,the lowerleft edges of BD 9174 match with the upper right edges of BD 1 1 957,and the right edge of BD9178 matches with the left edge of the patched BD 9174＋BD 1 1 957 segment.What is more,allthe incomplete characters on the edges can be matched.So it can be determined that BD 9174,BD1 1 957 and BD 9178 can be patched up.3.The same script.Dunhuang manuscripts had different scripts.Most of them are Regularscript,while some are Seal script,Clerical script,Cursive script or Running script.In addition,there is a kind of script,called″L-iKai,″which has the characteristics of both Clerical script andRegular script.The same script can also help us decide whether two fragments can be patchedup.For example,the scripts of Дх.100 and BD 5935 are both″L-iKai.″This information is clearevidence that Дх.100 and BD 5935 were split from the same document.4.The same writing style.Usually,documents copied by the same scribe in the same periodhad the same writing style,including the characteristics of the strokes and the structures of thecharacters.Thus,comparing the writing styles in different fragments is helpful.For example,by comparing the characters安、受、解、开、形、于、物、此、越、淫、轻 used in both BD 3153 and BD2852-1,we have found that they were written in the same style.It confirms that Дх.100 and BD5935 can be patched up.5.The same scribe.Some Dunhuang manuscripts have signatures of scribes.If differentfragments with connected contents have the signatures of the same scribe,they may be parts ofthe same document.For example,BD 3482 has three signatures of″rama？aHongzhen″in theseam of two sheets on the reverse side,while S.6440 has the same signatures on both the frontand the reverse sides.All of the signatures had the same style as the characters in the content,which suggests that Hongzhen was the same scribe of both BD 3482 and S.6440.The signature played a major role in patching BD 3482 and S.6440 up.6.The same reciter.Like the scribe＇s signature,the signatures in Dunhuang manuscriptsleft by reciters can also help in patching-up.For example,both S.3526 and BD 14031 have thesignatures of″Yizhen,″while the writing style is different from that of the characters in the bodyof the scripture,which suggests that Yizhen was the reciter rather than the scribe and he recitedboth S.3526 and BD 14031.The signatures help to determine that S.3526 and BD 14031 can be patched up.7.The same collector.Private collectors might deliberately tear Dunhuang manuscripts upinto several fragments.Thus the information about the collectors preserved in the fragments isimportant.For example,both ZD 170 and ZD 171 used to be collected by Zhang Zongxiang,andthey share the same serial number 23280 · 21 in Zhejiang Museum,which means that they arecollected as two fragments of one document.As expected,it turned out that ZD 170 and ZD 171 were torn apart from the same document.8.The same format.The manuscripts of Buddhist scriptures from Dunhuang have a standardformat which is,generally speaking,26 cm high,48 cm wide,25 columns per sheet in theSouthern and Northern Dynasties（420—589 CE）,and 28 lines per sheet in the Sui（581—618CE）and the Tang（618—907 CE）Dynasties.17 characters per column is the norm.However,BD 9187,BD 9188 A and BD 7925 are the only three mini fragments that are about 14 cm high,and contain 12 characters per column among 378 DunhuangBayang J ingmanuscripts.It canthus be inferred that they were split from the same mini document.9.The same correction or annotation marks.Buddhist scriptures would be proofread afterbeing copied.So a number of corrections were left either by the scribe or by the proofreader inDunhuang manuscripts, such as inserting the missing characters, deleting the redundantcharacters,crossing out the wrong characters,and reversing the flipped characters.To helppeople understand the context, reciters might also add annotations, such as segmentation,punctuation,and phonetic notation,etc.The same correction or annotation marks can help withthe patching-up.For example,DB 6 and DY 374 both have the reversal mark″ ″,which meansto reverse the characters before and after it,and all the reversal marks are of the same shape.Itshows that these two fragments can be patched up.10.Similar stains or damages.Documents may be stained or damaged before being splitapart.When the stained or damaged documents were split,the stains or damages in differentfragments often have similar shapes and similar colors.For example,a large portion of the stainsin BD 2556 and BD 2438 look the same and reappear regularly.These regular stains show thatthey were from the same document.1 1.The same version.One scripture may have several different versions.For example,theDiamond Sūtra has six translations,and the one translated by Kumārajīva gave rise to moreversions.One version has 32 sections,one has 12 sections,and another was added 62 words.Different versions are mostly identical but have minor differences.Comparing the differences canhelp with the patching-up.For example,BD 1404 and JY 213 both have 32 sections,which isrelatively rare among Dunhuang manuscripts.This is a clear clue that they can be patched up.12.The same binding format.The scroll is the mainstream and typical binding format ofDunhuang manuscripts. Other special binding formats includeFanj ia（梵 夹）, whirlwind,concertina,booklet,butterfly,etc.These special binding formats are so uncommon that they areeasily noticed.For example,S.5443 and S.5534 were both copied in booklet format,and areboth 1 1cm high and 1 6cm wide.Not only the binding format but also the size of the format givesclear clues to the connections between S.5443 and S.5534.Although the 12 key factors above are introduced with 12 different examples respectively,in practice,they should be considered in combination with other factors so that the patched-upversions are reliable and well grounded.