Background of the Text

Working as part of the famous Sino-Swedish expedition, Folke Bergman (1902–1946) visited the deserts of what is now western and northwestern China and Mongolia a number of times between 1927 and 1934. He recovered some 10,000 writing strips in the area of the Ejina 額濟納 River, which are known as the Juyan Han strips 居延漢簡. About half of those strips came from one site in modern Gansu, Pochengzi 破城子, called Mu-durbeljin in Mongolian. Bergman visited Pochengzi three times. Two of those visits were brief: one when Bergman was first traveling along the line of watchtowers marking Han-era defenses in the area, the other when he passed by during his final journey through the area in 1934. Bergman made the initial excavations at Pochengzi during his only extended stay there, his second visit. He arrived on 27 December 1930 and remained until the harsh conditions drove him away on 25 January 1931. What had initially seemed an unprepossessing mound turned out to be a Han dynasty border post and the most productive site that Bergman explored during his travels. His recovery of artifacts and manuscripts by the thousands from the Han structures there was the first part of this archaeological site's story.Footnote ³

Decades later, between 1972 and 1976, Chinese archaeologists made further excavations at Pochengzi. Their work was much more extensive than Bergman's. They excavated the central site, which is defined by a surrounding wall of packed earth that is roughly square at approximately 47.5 by 45.5 meters and 1.8 to 2 meters thick. This wall shows signs of having been repaired in ancient times. The site also contains a smaller fortified tower (zhang 障) on the northwest corner of the main enclosure, 23.3 meters per side, with adobe brick walls 4–4.5 meters thick. Within the walled enclosure are thirty-seven rooms, all having packed earth walls and floors of mixed mud and grass. The archaeologists recovered 3,434 wooden writing strips from the rooms and pits they dug within the area of the walled enclosure and tower, most from the northern part of the walled area.Footnote ⁴ These included “Evaluating Swords.” Scholars generally call the manuscripts recovered during this second round of excavations the “new Juyan strips” (Juyan xin jian 居延新簡)—even today, when they are no longer particularly new. Still later finds in the same general region around the Ejina River from 1998 to 2002 are known as the Ejina Han strips 額濟納漢簡.Footnote ⁵

Archaeologists found the strips containing “Evaluating Swords” in the northwestern part of the enclosure, in a pit that is referred to by the abbreviation E.P.T. 40. (E stands for Ejina, P for Pochengzi, and T for tanfang 探方, a numbered section of the excavation site, in this case 40.)Footnote ⁶ Although “Evaluating Swords” has no date upon it, other materials from the same pit bear dates ranging from 9 b.c.e. to 21 c.e., giving a rough idea of when this copy of the text was buried.Footnote ⁷ This provides a general dating of the manuscript to the late Western Han, Xin, or early Eastern Han period.

The authors of the first transcription of “Evaluating Swords” make a very important distinction between the date of the manuscript from Pochengzi and that of the text's creation.Footnote ⁸ Based upon the copy's position and the layers of its location, presumably (although tacitly) with reference to the dated materials, they date the manuscript to late Wang Mang 王莽 (9–23 c.e.) or early Eastern Han (25–220 c.e.) times. They believe the content to date to the early Western Han.Footnote ⁹

The question of the content's specific time of creation is impossible to answer definitively, although a tentative general dating is possible. In his discussion of sword evaluation, Ma Mingda 馬明達 refers to the time from the Chunqiu 春秋 period to the Han as the “golden age” of swords in China. He attributes this to technological advances during the time, the spread of the study of sword theory and technique, and the function of swords as signs of social rank. Together with these things came connoisseurship, collecting, and high prices for rarities. That led in turn to the potential for counterfeiting valuable swords and the need for skilled assessment of examples. Lüshi chunqiu 呂氏春秋, from the late third century b.c.e., for instance, mentions this kind of identification as part of evaluating swords in considering the dangers of false semblances: “A sword evaluator worries about a sword's resembling Wugan” 相劍者之所患, 患劍之似吳干者 without being that famous blade.Footnote ¹⁰ These things are part of the intellectual background of “Evaluating Swords.”Footnote ¹¹ It must be noted, however, that nowhere does “Evaluating Swords” show any interest in recognizing famous swords or avoiding false identifications. It is concerned only with judging the quality of particular examples.

Ma Mingda furthermore sees indications within “Evaluating Swords” that its content dates to early Western Han times. First, and as others have also noted, its descriptions apply to swords made from worked and treated steel, not bronze or plain iron. This means the text must date no earlier than the Western Han, when steel became important in China.Footnote ¹² According to Ma and others, there was moreover a shift in the style of sword around the middle of the Western Han. Whereas formerly a straight sword with a knob pommel called the jian 劍 had been most popular, around the middle of the Western Han the dao 刀 sabre,Footnote ¹³ a sword with a blade that was usually slightly curved and had a ring at the pommel, supplanted it. Both terms appear in “Evaluating Swords,” but the latter just once, while the former occurs repeatedly. This indicates the dao had not replaced the jian when this text was created. On the basis of these two things, Ma dates the content of “Evaluating Swords” to the first part of the Western Han.Footnote ¹⁴

There is the final question of whether this manuscript text is a standalone composition or an extract from a larger work. Liu Jinhua 劉金華 proposes three possible origins: a late Western Han selection from earlier, sword-related texts; a record of contemporary experience with swords; or a selection from some other, no longer extant, contemporary sword text.Footnote ¹⁵ Along similar lines, Ma Mingda suggests it might either record the experience of a sword evaluator or be an extract from a larger text on the appraisal of swords.Footnote ¹⁶

Many readers have proposed that “Evaluating Swords” has a relationship with the title Xiang bao jian dao 相寶劍刀, which appears in the Han shu 漢書 “Yiwenzhi” 藝文志.Footnote ¹⁷ Li Ling 李零, for instance, uses the title from “Yiwenzhi” to denote the Juyan text.Footnote ¹⁸ Zhong Shaoyi 鐘少異 brings the title from “Yiwenzhi” into his discussion of dating “Evaluating Swords.”Footnote ¹⁹ Han Hua 韓華 and Di Xiaoxia 狄曉霞 blame the dearth of scholarship concerning “Evaluating Swords” on the fact that Xiang bao jian dao, listed in “Yiwenzhi,” is no longer extant.Footnote ²⁰ This appears to assume the two were the same or closely related, as only then would that comparison have been a necessary and productive enterprise.

Despite the variety of opinions scholars have expressed about its derivation, there is no evidence connecting “Evaluating Swords” to any other text. The intelligibility of “Evaluating Swords” as it is, and the lack of any indication that it is an extract, suggest that it is simplest and thus most preferable to treat it as an independent work until evidence emerges to the contrary. This is the approach I follow. Since there is no evidence whatsoever of a relationship with the title Xiang bao jian dao from the Han shu “Yiwen zhi,” the complete absence of the term xiang 相 (“to evaluate visually, to physiognomize”) underscores that any direct connection between the two is, at best, speculative. In the absence of evidence, asserting a connection between that title and this text is thus more likely to confuse the situation than clarify it.

The State of the Manuscript and Its Text

The hand that copied out the “Evaluating Swords” manuscript did so in a clear and legible clerical script and the strips we have are very well preserved.Footnote ²¹ The result is a text that, with the exception of a few graphs, is fairly readable in photographs (see illustration 1). Most—but not all—problems in the interpretation of the text stem from sentence structure and vocabulary, not handwriting. And there are some significant difficulties in that respect, as I will discuss below.

Illustration 1. (Hu Zhi 胡之, ed., Neimenggu Juyan Han jian (er) 內蒙古居延漢簡(二) [Chongqing: Chongqing, 2008], 13.)

“Evaluating Swords” as we have it comprises six wooden strips (pine), approximately 22.6 cm long and 1.2 cm wide. Each strip contains one line of text, giving a total length of 209 graphs. There are seven bullet points in the text, marking breaks of different sorts; no other punctuation occurs.Footnote ²² There are notches on the strips for three lines of binding, one each at the top and bottom and one in the middle of the strips, but the bindings are not present and the strips are loose.Footnote ²³ The separation of the strips invites questions about the sequence and integrity of the text as we have it. Due to the flow of the text and the presence of brief summaries of content in it (see below), there have not been significant questions about the sequence of the strips. The completeness of the text is another matter.

The brief introduction to the first published transcription of “Evaluating Swords” asserts that content is missing between the penultimate and final strips. The authors of that transcription posit a discontinuity in the prose at that point, which they believe indicates the absence of at least one strip. Many researchers have accepted this postulation, without speculating about the nature of the missing text. Opinions about the amount of material that was lost have differed.Footnote ²⁴

Not all who have written about “Evaluating Swords” agree that there is a significant break in its text. Liu Jinhua and He Maohuo 何茂活, for instance, separately argue on the basis of structure, textual parallels, and content that the only thing missing is one of the summaries within the text, and not any significant content. He Maohuo suggests the summary in question was, by accident or not, left out when copying and that no strip was lost.Footnote ²⁵

We cannot know with certainty whether “Evaluating Swords” as we have it is complete. This is the nature of writing strips: once they have come unbound, and absent other indications, pretty much anything is possible. Yet the text as it comes to us is intelligible, and I agree with Liu and He—contra other opinions—that there is no apparent lacuna in the main content. At the same time, there is some material that does not fit into any of the summaries present in the text, suggesting at least one more such summary should exist. Thus, the most parsimonious explanation is that a summary is missing from this copy of the text and nothing else. This essentially matches the proposals of Liu Jinhua and He Maohuo. It is of course impossible to prove this on the basis of present evidence, and even if proven this would say nothing about the origin of the content or its original state. Ma Mingda's arguments and the archaeological context provide a fair indication that it must come from a time around the middle of the Western Han. The dated materials from the same pit leave us roughly certain that the copy “Evaluating Swords” we now have was in use around the time 9 b.c.e. to 21 c.e.

The Content of “Evaluating Swords”

In this section, I briefly summarize and discuss the content of “Evaluating Swords.” In an attempt at simplicity, I will refer to the text of “Evaluating Swords” by strip sequence and serial numbers, without referring to the transcriptions or otherwise adding notes except as necessary for the discussion here. I explain the interpretive details of difficult words and phrases in the section that follows this one. For citation information and further notes, please see the transcription and translation at the end of this article.

As I mentioned at the outset, “Evaluating Swords” begins with a bullet point and the condition, “If you want to know whether a sword is good and old.” The first step in the process to reach that determination follows that condition, which is the direction to “stand and draw it” 起拔之. This reflects that the text bases its evaluations on the blade, not on the pommel, grip, scabbard, or intangible elements.

Since the opening of “Evaluating Swords” refers to both “good” and “old,” some scholars have taken these two characteristics in tandem, explaining the text as expressing a simple preference for antiquity that ranks any old sword over any new one because of its value as a mark of social status and as a collector's item.Footnote ²⁶ As Ma Mingda points out, however, a careful reading of “Evaluating Swords” shows it to be practical in focus. For a sword to be old is certainly in its favor; I think this is probably because its quality is proven. But “Evaluating Swords” makes no mention of rarity or special monetary value; nor does it refer to any embellishment, such as one might expect to grace an implement made for show more than for use. Its descriptions concern only things that indicate properties of the blade's metal. This text is not about how to find a collector's item or an antique but rather how to identify a quality sword for use.Footnote ²⁷ There is also no information anywhere in the text about why a particular characteristic indicates good quality, only the insistence that it does so.

The process for identifying an old sword according to “Evaluating Swords” consists of one step: seeing that the main part of the blade lacks remaining marks from metalworking. (It calls these tuichu 推處, a term I discuss below.) After this, “Evaluating Swords” drops the question of whether a given sword is old, going on to discuss three other characteristics indicating good quality. We know there are four characteristics in total identifying a good-quality sword because the fourth strip (E.P.T. 40:205) refers to that number of characteristics in its summary of the section. This indicates that whether a sword is old or not is the first of the four characteristics treated in the opening section of “Evaluating Swords.”

The other three good qualities concern the quality of the metal as observable from a visual inspection. “Evaluating Swords” prescribes looking at a blade for the presence of “paired black lines that are unbroken and that at the tip these seem to disappear” 黑兩桁不絕者, 其鋒如不見. This apparently describes a blade that a smith made by fusing two different types of metal together, one light in color and one dark, to give a blade that balances resilience with hardness. High carbon ferrous metal is light in color and, while very hard, also brittle. Lower carbon ferrous metal is darker; it is softer but also tougher and less vulnerable to breaking. The welding together of different metals to achieve a result that balanced the strengths and weaknesses of the two existed in China no later than the third century b.c.e.Footnote ²⁸ This layering seems to be connected to what the text denotes as the “white and hard” 白堅, though it does not use the term here.

The third matter of a good sword concerns the “white and hard”—one of the two types of metal I have just discussed—which “Evaluating Swords” says should end before coming to the final third of the blade. He Maohuo and Ma Mingda cite Lüshi chunqiu in this context.Footnote ²⁹ There we find an anonymous evaluator of swords, who explains the makeup of a two-color blade: “The white is that which makes it hard; the yellow is that which makes it tough. If the yellow and the white are mingled, it is both hard and tough, a fine sword” 白所以為堅也, 黃所以為牣也. 黃白雜則堅且牣, 良劍也.Footnote ³⁰ Aside from the difference in color describing the durable metal component—which might result from technological differences, variation in adjective usage, or simply perceptions—the Lüshi chunqiu passage could tally with what we see in “Evaluating Swords” in terms of a two-tone blade. This reading of the Lüshi chunqiu, however, does not accord with the one Joseph Needham presented in his study of ferrous metal technology. He followed Richard Wilhelm's translation of the Lüshi chunqiu, which says that the two kinds of metal are “gemischt” (mixed). Needham thus takes the Lüshi chunqiu passage to describe an alloy rather than welded layers, which then would not fit the descriptions of blades in “Evaluating Swords.”Footnote ³¹

What is perhaps unexpected in “Evaluating Swords” is that the lighter-colored material is supposed to appear only in the lower two-thirds of the blade, with the final section, at the tip, made up of the darker material only. That the text treats this as a positive trait is made doubly clear below, when it cites the opposite—the two colors extending to the tip—as a negative point. A sword with a tip made up of darker metal only is described as one of the best in the realm, though not yet perfect or even of the highest level. The next sentence, however, names the last of the four positive characteristics, the presence of an “appearance like millet grains” 如黍粟狀 on the body of the blade, which makes a sword even better. The sense of the text here is clear but the specific significance is obscure. No one who has written on the subject has been able to refer to an instance of this or similar description in an early text, with the predictable result that readers have explained it in different ways. Liu Jinhua, for instance, conjectures that this describes a layer of oxidation with a spotty appearance, which developed on the blade after a special treatment and protected it from decay.Footnote ³² Ma Mingda refers to stories of famous swords that are supposed to have had patterned blades. He says that this particular pattern resulted from the wrought steel construction of the blade, which as a result developed a spotted layer of oxidation on the body (as opposed to the edge).Footnote ³³ Chen Li 陳力 explains this as describing the places where the dark and the light metals come together, which would be of a different appearance on the body of the sword than on its edge.Footnote ³⁴ I find Chen Li's explanation most persuasive because it continues the focus on the intrinsic quality of the metal itself, rather than introducing new factors. Chen's explanation furthermore tallies well with images of wootz and damascene blades that I find online. But there is at present no certainty about what exactly might have created a millet grain appearance on a sword. This ends the section of Part 1 that discusses good characteristics of swords and the second strip.

“Evaluating Swords” next turns to the characteristics of poor swords. Like the preceding section, which dealt with good swords, this one begins with a bullet point. While the bullet point marks a break, the summary at the end of the fourth strip (E.P.T. 40:205) implies that everything preceding it, regarding both good and poor swords, forms one section. I thus treat this also as a portion of Part 1.

This section begins, like the previous does, by addressing and defining its audience: “• One who wishes to know if a sword is poor and so not to use it, or that it is new equipment” • 欲知弊劒以不报者, 及新器者. This brings out the distinction between a poor sword and a new one by means of an explicit “and” (ji 及), making the separation even stronger than the opening of “Evaluating Swords” does when it distinguishes good swords and old ones. This separates new swords from those necessarily bad, which leads me to think that the objection to a new sword is less absolute than to a bad one. In conjunction with the preference for old equipment I have already discussed, I suggest this advice is intended to help the one evaluating a sword for use avoid an untested blade in favor of something of proven reliability.

“Evaluating Swords” begins its discussion of bad swords and new ones with the command to go into the sun and make a careful visual inspection. Just as the lines created by layers of lighter and darker metal were supposed to disappear before the tip of a good sword, so is their continuation to the tip the mark of a poor one. Similarly, the presence of gouging marks a bad sword, just as the absence of such marking was characteristic of a good one. The dark and light metal should also not be “clearly separated,” perhaps because that would reflect an imperfect or incomplete welding process.

Finally, the text discusses patterning on a blade. Chen Li understands this to refer to patterns created by processes of hardening and tempering the blade; he believes the text has shifted its focus from the metal intrinsic to a blade to changes in appearance due to those processes.Footnote ³⁵ There is no doubt that early metal- and swordsmiths in China knew and employed techniques of hardening and tempering, including quenching in a variety of liquids. Processes like those can indeed leave patterns on the surface a metal object.Footnote ³⁶ In the present case, however, I believe the patterns discussed are more likely to be those of the metal itself and not those left by quenching or something else, for two reasons. First is the absence of any indication of a shift in topic, without which continuation of the preceding topic is likely. Second is the conclusion of this subsection, which says that the preceding characteristics are those of “swords from poor welders,” suggesting again that the text continues the focus on the formation of the weapon—i.e., the work of the smith.

In this section, patterns indicate poor quality in several ways. One way is when the patterns follow the length of the sword in the hard portion of the blade. A continuous swirling pattern also reflects a poor sword. This section says in closing that its provisions apply equally to swords (jian) and sabres (dao). This is the only place the latter term occurs in this text; this seems like it could be a later addition to a previously existing text about swords (jian) to make explicit that its advice applies also to sabres. This section closes with the internal summary giving the numbers of matters of good swords and bad that it treats: four and six, respectively.

The final part of “Evaluating Swords” talks about specific patterns and final characteristics of blades that are desirable and those that are not. The patterns are described in an impressionistic manner, the preferable ones being “Hanging Curtains,” “Bearing Paired Snakes,” “Holding Feathers,” and “Jade Tablet”; those to avoid are “Fighting Cocks” and “Writhing Snakes.” Obviously it is impossible to know what these mean specifically but the general sense is clear.

This last section also reiterates that the focus of “Evaluating Swords” is on appearance, asking rhetorically, “If a strong one could have a bad exterior appearance, what would a weak one be like?” and noting that a coarse appearance is a sign of bad quality. The entire thing ends with a final summary, which mentions only the four qualities of bad swords.

The Structure of the Text

Bullet points appear in “Evaluating Swords” at the top of the first strip and three others (E.P.T. 40:202, E.P.T. 40:204, E.P.T. 40:206, and E.P.T. 40:207), in each case beginning a short section concerning one type of judgment. Bullet points are widely seen in excavated materials and often serve to signal the start of a sentence or section or to mark another sort of break. Establishing breaks like this is helpful for the reader; even more useful is an indication of how its segments group together, and the internal summaries in “Evaluating Swords” serve this purpose.

At the end of the third strip (E.P.T. 40:205) comes “• The preceding are four matters of good swords. • The preceding are six matters of bad swords” • 右善劒四事 • 右弊劒六事. These two labels have bullet points that divide them from the preceding prose text and from each other. Most useful, in terms of analyzing the structure of the text, is the information they provide about the quantity of decision criteria, which helps parse the material and establishes the first main division of the text.Footnote ³⁷

Due to the absence of other punctuation, the number of points of evaluation for a good sword would be at least potentially ambiguous without the summary's indication that there are four. Knowing that there are four, however, suggests that the occurrence of four instances of the word shi 視, “to look,” at the beginning of sentences describing characteristics of good swords is more than just repetition. He Maohuo notes the pattern and argues that it marks a parallel structure, with each of four sentences beginning shi, in one case (second strip, E.P.T. 40:203) prefaced with you 又, “also, again.” Although not all scholars accept this division, the symmetry of four instances and four characteristics, as well as the parallel structure of the sentences, indicates to me that this is the best approach.Footnote ³⁸ The closing words of “Evaluating Swords” form a final summary that encapsulates not the entire text but only the last four points as “matters of poor swords.”

Problems of Interpretation

There are several terms and phrases that pose difficulties in reading “Evaluating Swords.”Footnote ³⁹ One of these is numbered among the decisive characteristics of a sword: the text refers to tuichu 推處, the absence of which indicates an old sword. This binome does not appear in this context in received texts. It has thus, inevitably, occasioned some interpretive creativity on the part of readers. Ma Mingda proposes that tuichu is a commoners’ term for the quillons of a sword. He says that Chunqiu-period bronze swords had either no crossguards or only very small ones, and as such their absence would mark an “old sword.” Ma admits that there is no support for his reading other than its sense in this context.Footnote ⁴⁰ A more important shortcoming is the fact that he does not link this with the rest of “Evaluating Swords” and its focus on steel swords and practical use. He Maohuo suggests instead that tuichu denotes places where the blade has been damaged by being struck by another blade. It is not obvious why the absence of such damage would indicate an old sword and he does not explain this point.Footnote ⁴¹ Li Ling proposes that tuichu denotes the division between two colors indicating different types of metal, without canonical support for his reading or explaining why it would identify the blade as an old one.Footnote ⁴²

Chen Li takes tuichu as marks in the metal left over from the process of making the sword. He offers two possible lines of reasoning to reach this conclusion. The first is the simplest, taking the components of tuichu in their usual senses to give “push places,” which he explains as “shoveling,” i.e., gouges in the metal left behind from initial working of the blade. Alternately, he suggests tui 推 could also be written here for chui 椎, “hammer; to hammer,” which would mean “hammer places,” i.e., hammer marks from forging that remained on the blade. Both suggestions seem to me very reasonable. However they came to be, in Chen's proposal the marks would have worn away over time through maintenance and use, their absence thus indicating age.Footnote ⁴³ While the lack of canonical examples of tuichu makes a final determination impossible at present, I concur with Liu Jinhua that Chen Li's explanation is the best that has been offered so far.Footnote ⁴⁴

The phrase I translate “paired black lines” is hei liang heng 黑兩桁. The color “black” (hei 黑) is clear without further clarification. Liang 兩, usually “two,” I understand following He Maohuo to function verbally, “to be two, to be paired.” Heng 桁 means “purlin(s),” horizontal roof rafters, and is here used metaphorically to refer to the appearance of the layers of metal.Footnote ⁴⁵ A blade that “seems to be invisible” 如不見 accompanies these lines in the metal. This last point likely refers to a keen edge in terms familiar to anyone who has successfully sharpened a knife.

When the text discusses the characteristics of a poor sword, it relates the result of reaching such a conclusion on the third strip (E.P.T. 40:204) with a phrase that is difficult to understand. All published transcriptions have the three graphs in question as yi bubao 以不報, and the photographs leave no doubt that this is the proper transcription of the first two characters. The identification of the third character (previously bao 報), however, gives cause for uncertainty. The phrase yi bubao is itself by all accounts difficult to understand, due mainly to the presence of the word that previous transcriptions unanimously read as bao 報, which usually means “to repay, requite, respond.” The idea of a sword that does not repay is not familiar from other texts. Some authors who have treated “Evaluating Swords” suggest that this repayment refers to bringing good luck. These authors do not support their reasoning with references to other texts. Presumably they have in mind usages like that occurring in the aphorism “Hidden virtue always has an evident reward (bao)” 陰德必有陽報. Versions of that saying appear in a number of Han texts, sometimes with “heaven” (tian 天) named as the source of the reward, but often not.Footnote ⁴⁶ The context in “Evaluating Swords,” however, concerns quality rather than luck, and the unusual nature of this interpretation with reference to a weapon leaves me doubtful.

Examination of the published photographs of this strip increases my doubts about this identification. Comparison of the graph generally transcribed as bao 報 with the examples listed in Lu Xixing's 陸錫興 Handai jiandu caozibian 漢代簡牘草字編 confirms that the right part of the graph is like that of bao in other contemporary manuscripts, namely forming fu 𠬝. The left-hand part, however, is different from that in other examples of bao, in that it has a hard angle leading to a long stroke to the left at the bottom (see illustration 2). This difference is not enough to make identification of the graph as bao impossible, and the unanimity of previous readers is, of course, not without reasonable basis. Yet the combination of an unusual semantic sense and an unusual graphic form makes me think another reading may be preferable.

Illustration 2. (colour online) (Hu Zhi, 13.)

I suggest that this graph is a previously unattested variant of fu 服. There are two parts to this identification. The first part is the easier and concerns usage. There are many examples of fu 服 used in the sense of “to bear (a sword).” The examples easiest to locate come in dictionaries like Ciyuan 辭源 and Hanyu dacidian 漢語大詞典, which list the phrase fujian 服劍, “to wear a sword, to bear a sword.”Footnote ⁴⁷ Moreover, the word fu 服 has, among its many senses in the dictionaries, the meanings “to use” 任用 and “to bear in the hand” 執持. Understanding yi bufu 以不服, “in order not to bear (that sword),” would therefore be a straightforward reading in this context.

The second part of the identification concerns the form of the graph. Shuowen jiezi 說文解字 says that fu 𠬝 is the phonetic element of the graph fu 服, which it glosses as “to use.”Footnote ⁴⁸ I suggest the left part of the graph in “Evaluating Swords” is an anomalous shou 手, “hand,” with an additional horizontal stroke, an attested manuscript form.Footnote ⁴⁹

Read in this fashion, a standardized modern version of the graph before us would be 报, read as fu and meaning “to bear, to use.” This character is of course in use in the People's Republic of China today as the simplified form of bao 報, which perhaps contributes to the more common identification of the character in transcriptions of “Evaluating Swords.” There are also broader semantic and phonological connections between bao and fu, which provide some incidental support for the relationship between the two.Footnote ⁵⁰ I suggest that in the Han manuscript context, it is permissible in terms of graphic form and preferable in sense to take the character to be 报, read in context as fu, “to bear,” which gives my translation.

In reading further along in the third strip (E.P.T. 40:204), I again propose an interpretation that is unlike any published opinion. Previous transcriptions have in the text there xing 騂, “brownish red,” usually describing the color of a horse or an ox. Ma Mingda suggests this denotes a yellowish-red color in the metal of the blade resulting from the oxidation of impurities left by inferior workmanship.Footnote ⁵¹ The problem with this reading is the position of the graph in the sentence, before the verb “to look at,” whereas in that sense one would expect it to follow the verb. Chen Li proposes reading xing as a loan for qie 挈, “to lift in the hand(s), to suspend,” giving the direction to “lift and look at it.” This is reasonable in terms of sense and the alternation between xing and qie is attested in other sources, yet the lack of phonetic or graphic similarity raises questions about a general interchangeability between the two.Footnote ⁵² Moreover, my examination of the photographs of the strip suggests that the previous transcription may be incorrect.Footnote ⁵³ At the very least, the character in question is not a standard version of xing. For while xing is composed of ma 馬 and xin 辛, the graph appears to be instead , comprised of the radical ma and the phonetic element da/ta 𦍒 (see illustration 3). This character is also unlike the example of xing that Lu Xixing lists from among Han manuscripts.Footnote ⁵⁴

Illustration 3. (colour online) (Hu Zhi, 13.)

As in the previous case, the published reading is certainly possible. Yet the combination of interpretive difficulty and graphic variation again leads me to think the best reading may differ from earlier suggestions, and I propose a novel graphic variation. The left-hand part of this graph, typically the semantic portion, is clearly ma 馬, “horse.” As noted already, the right-hand portion, apparently the phonetic element, is da/ta 𦍒. On the basis of the shared phonetic element I suggest the graph is a graphic variant or loan for the word da 達, “to reach, achieve.” There is a conceptual relationship between the semantic elements of the two variants. Chuo 辵/辶, the radical of da 達, often indicates a connection with movement or travel; ma as a semantic element can do the same for equestrian travel and sometimes by extension for other forms of travel, too.Footnote ⁵⁵ Da furthermore works well in terms of sense here, as together with the next graph it forms the binome dashi 達視, “to look at carefully,” which is attested in received Han texts and logical in context.Footnote ⁵⁶

After listing various negative characteristics of swords, the fourth strip (E.P.T. 40:205) says that all are bi he ren 幣(:弊)合人 swords. The content of the preceding strip and this one (i.e., E.P.T. 40:204 and E.P.T. 40:205) flows together seamlessly, producing a list of six characteristics. And as I noted above, strip four (E.P.T. 40:205) ends by saying, “The preceding are six matters of poor swords.” This last point makes it certain that the phrase bi he ren refers to the preceding poor swords. Bi 幣(:弊) occurs repeatedly in the text in the sense of “bad” and ren 人, “person,” seems obvious enough. But the three words together leave the sense of the word he 合, usually “together; to meet together; to close,” difficult to comprehend. He Maohuo, for instance, acknowledges the difficulty and proposes understanding the phrase as referring to swords that are of poor quality and yet can “go with (he) a person.” He explains this as describing swords that are not “good” but are also not unpropitious.Footnote ⁵⁷ Yet the general practical and concrete focus of “Evaluating Swords” makes me doubt this reading, to say nothing of the unusual usage and the non-standard grammatical structure it demands.

The word he 合 can refer to bringing together or combining things. Thus, Chen Li suggests that here it means metal “alloy.” Chen explains away “person” (ren 人) in this line by calling it an error for the grammatical particle zhi 之. In his reading, the phrase is bi he zhi jian 弊合之劍, “a sword of poor alloy.”Footnote ⁵⁸ While attractive in terms of sense, the problem with this interpretation is that the character for “person” in this text is clear and not similar to that of zhi, which appears repeatedly elsewhere in “Evaluating Swords.” Nevertheless, I think Chen is correct to understand this phrase as denoting, in some way, the material quality of the blade.

Chen Li's interpretation informs mine but my understanding differs from his in two main respects. First, since the preceding phrases refer to patterns in the blade resulting from layers of metal, I take he to mean not “to alloy” but rather “to weld”Footnote ⁵⁹ together, i.e., the process of forming the blade from multiple pieces of metal, resulting in the layering already discussed. I take “person” (ren) as written. Together, I suggest, the two words heren 合人 refer to the “welder,” i.e., the person who carried out that task in the manufacture of the blade. Zhouli 周禮, probably dating to the late Warring States period, contains numerous examples of titles comprising a noun or verb plus “person.”Footnote ⁶⁰ In my reading, bi heren jian means “swords from poor welders,” which were low-quality blades exhibiting characteristic patterns that a buyer should avoid.

Translation of “Evaluating Swords”

My transcription, punctuation, and translation draws from a number of previously published versions and related research articles.Footnote ⁶¹ The transcription as reproduced here as a rule employs standardized modern versions of all characters, with a few graphs that are not part of standard character sets. As I mentioned in the introduction, the original strips are not numbered and contain no punctuation other than bullet points. I insert sequential numbers and include also the archaeological serial numbers for ease of reference. In the translation I insert the sequential numbers in square brackets at the corresponding places.

1. 1 • 欲知劒利善故器者, 起拔之, 視之身中無推處者, 故器也. 視欲知利善者, 必視之身中有黑兩桁不絕者, (E.P.T. 40:202)

2. 2 其逢(:鋒)Footnote ⁶² 如不見. 視白堅未至逢(:鋒)三分所而絕, 此天下利善劒也. 又視之身中生如黍粟狀, 利劒也, 加以善. (E.P.T. 40:203)

3. 3 • 欲知幣(:弊)Footnote ⁶³ 劒以不报(:服)Footnote ⁶⁴ 者, 及新器者, 之日中, (:達)Footnote ⁶⁵ 視白堅隨逢Footnote ⁶⁶ 上者及推處、白黑堅分眀者, 及無文、縱有 (E.P.T. 40:204)

4. 4 文而在堅中者, 及雲氣相遂, 皆幣(:弊)合人劒也. 刀與劒同等 • 右善劒四事 • 右幣(:弊)劒六事. (E.P.T. 40:205)

5. 5 • 利善劒文: 縣(:懸)薄文者保雙虵文皆可. 帶羽圭中文者皆可. 劒 (:鳴)Footnote ⁶⁷ 者利善. 強者表𦷚(:惡),Footnote ⁶⁸ 弱則利奈何? (E.P.T. 40:206)

6. 6 • 𦷚(:惡)新器劒文鬬雞佂虵文者, 麤者及皆凶不利者 • 右幣(:弊)劒文四事 (E.P.T. 40:207)

[1] • One who wishes to know whether a sword is good and old should stand and draw it. If you look at it and the body has no gouging, it is old equipment. One who, looking at [a sword] wishes to know if it is a good one, should always look to see that in its body there are paired black lines that are unbroken [2] and that at the tip these seem to disappear. If you look and see that the white and hard ends before reaching the last third, this is a good sword of the realm. If you see also that the body has in it an appearance like millet grains and it is sharp, it is even better.

[3] • One who wishes to know if a sword is poor and so not to bear it,Footnote ⁶⁹ or that it is new equipment, should go into the sunlight. If, when you look carefully,Footnote ⁷⁰ the white and hard follows up the tip; or there is gouging, or the black and the hard are clearly separated; or there is no patterning, or along the length there is [4] patterning and it is in the hard; or the clouds and ether (i.e., swirling patterns) follow each other; then these are all swords from poor welders.Footnote ⁷¹ Sabres and swords are of the same sort.

• The preceding are four matters of good swords. • The preceding are six matters of bad swords.

[5] • Patterns on a good sword: “Hanging curtains” and “Bearing paired snakes” are both acceptable. “Holding feathers” and “Jade tablet” are both acceptable. A sword that sings is good. If a strong one could have a bad exterior appearance, what would a weak one be like?

[6] Bad new equipment with the sword [blade] patterns of “Fighting cocks” or “Writhing snakes,” and those of coarse make, are all unpropitious and poor. • The preceding are four matters of poor swords.