Evidence from texts — examining characters and books書證

The Odes say, "The ragged heng plants (荇菜)." The Erya says, "Heng is jieyu (接余)." The character is sometimes written heng (莕). The explanations of the earlier scholars all call it a water plant, with round leaves and a slender stem that follows the depth of the water. Today it is found wherever there is water. Its yellow flower is like that of the water shield, so the common speech of Jiangnan calls it pig water shield, and some call it hengcai. Liu Fang annotated it fully. But the common people of Hebei mostly do not know it, and take what the erudites all call "ragged" to be amaranth, calling the amaranth that people eat "human heng." Nothing could be more laughable.

The Odes say, "Who says that tu (荼) is bitter?" The Erya and the Mao Commentary on the Odes both take tu to be the bitter herb. The Book of Rites also says, "The bitter herb puts out its ears." I have looked into it. The Yi tong tong guayan xuantu says, "The bitter herb comes up in the cold autumn, passes the winter, goes through the spring, and only reaches its growth in summer." The bitter herb of the Central Plain today is like this. It has another name, youdong (遊冬). Its leaves are like those of the sow thistle but finer; broken off, it gives a white sap; its flower is yellow, like a chrysanthemum. In Jiangnan there is a different bitter herb, with leaves like the ground cherry, flowers either purple or white, and fruit as big as a bead, red or black when ripe. This herb can ease weariness. I have looked into it. In Guo Pu's note on the Erya this is the zhi huangchu (蘵黃蒢), what Hebei now calls longkui (龍葵). Those who expounded the Book of Rites under the Liang took this for the bitter herb; but it has no old root and only comes up in spring, so that too is a great mistake. Again, Gao You notes on the Lüshi chunqiu, "What flowers but does not fruit is called ying (英)." The bitter herb ought to be called ying, and from this it is all the clearer that it is not longkui.

The Odes say, "You di zhi du (有杕之杜)." The Jiangnan copies all write the tree radical with da (大) beside it, and the Commentary says, "Di is the look of standing alone." Xu Xianmin gave the sound as the fanqie of tu (徒) and ji (計). The Shuowen says, "Di is the look of a tree," and puts it under the tree radical. The Yunji gives it the sound of the di (第) of cidi (次第). But the Hebei copies all write the di (狄) of yidi (夷狄), and read it accordingly as that character. This is a great mistake.

The Odes say, "Jiongjiong mu ma (駉駉牡馬)." The books of Jiangnan all write the mu (牡) of pinmu, female and male; the Hebei copies all write the mu (牧) of fangmu, putting out to pasture. An erudite of Ye raised an objection: "The Jiong ode praises Duke Xi for pasturing horses in the open country of Jiong. Why should it be held to mares or stallions?" I answered: "I have looked into it. The Mao Commentary says, 'Jiongjiong is the belly and body of a fine horse, fat and full.' Below that it says again, 'A feudal lord has six stalls with four kinds: fine horses, war horses, field horses and draught horses.' If it were taken in the sense of pasturing and ran over both female and male, then the name jiongjiong could not be held to fine horses alone. Fine horses draw the Son of Heaven's jade carriage, and serve the feudal lords for court visits and for the suburban sacrifice; there are certainly no mares among them. In the Zhou li, under the office of the groom, 'For fine horses, one man to a horse; for draught horses, one man to a pair.' What the groom rears is not mares either. The maker of the ode brought forward the strong and the swift, and that suits the sense. The Yi says, 'The fine horses gallop.' The Zuo zhuan says, 'With his two fine horses.' These too are names for the choice and the swift, not words of general use. To make the fine horses of the Commentary on the Odes run over mares at pasture is, I fear, to lose Master Mao's meaning. And have you not seen Liu Fang's Yizheng?"

The Yueling says, "Liting (荔挺) comes out." Zheng Xuan's note says, "Liting is the horse-leek." The Shuowen says, "Li is like the cattail but smaller, and its root can be made into a brush." The Guangya says, "The horse-leek is li." The Tongsu wen also calls it malin (馬藺). The Yi tong tong guayan xuantu says, "If liting does not come out, there will be many fires in the state." Cai Yong's Yueling zhangju says, "Li is like ting (挺)." Gao You's note on the Lüshi chunqiu says, "The li plant thrusts straight up." If so, the note to the Yueling was wrong to take liting for the name of a plant. It grows freely in the flat marshes of Hebei. There is a good deal of it in Jiangdong too, and people sometimes plant it in their courtyards, but they call it only handpu (旱蒲) and so do not know it for the horse-leek. Those who expound the Book of Rites take it for maxian (馬莧). Maxian can be eaten; it is also called tun'er (豚耳), and its common name is machi (馬齒). At Jiangling there was once a monk whose face was broad above and narrow below. Liu Huan's young son Minyu, who was only a few years old and was quick-witted and good at catching the look of things, saw this monk and said, "His face is like maxian." The boy's uncle Liu Tao thereupon called the monk the Liting Master. Liu Tao was a noted scholar who himself expounded the Book of Rites, and even so the mistake went as far as that.

The Odes say, "He comes shishi (施施)." The Mao Commentary says, "Shishi is the sense of coming forward with difficulty." Zheng's gloss says, "Shishi is the look of walking at ease." The Han Odes also doubles it as shishi, and the Mao Odes of Hebei all give shishi. The old copies of Jiangnan all have the single shi, and the common run have come to hold that right; I fear it is a small mistake.

The Odes say, "The dark clouds are thick, the clouds rise qiqi (有渰萋萋，興雲祁祁)." The Mao Commentary says, "Yan (渰) is the look of dark cloud. Qiqi (萋萋) is the look of cloud moving. Qiqi (祁祁) is the look of slowness." The gloss says, "In old times yin and yang were in harmony and wind and rain came in season, and their coming was qiqi and not violent." I have looked into it. Yan is already dark cloud; why trouble to say again "the clouds rise qiqi"? Yun (雲) should be yu (雨); the common copyists have got it wrong. Ban Gu's poem on the Spirit Terrace says, "The three lights spread their brightness, the five agents set out their order; soft blows the auspicious wind, qiqi falls the sweet rain." This is the proof.

The Book of Rites says, "Settle what is youyu (猶豫), decide what is doubtful." The Li sao says, "The heart is youyu and huyi (狐疑)." None of the earlier scholars has explained them. I have looked into it. The Shizi says, "A dog five feet long is a you (猶)." The Shuowen says, "In Longxi a puppy is called you." To my mind, when a man walks with a dog, the dog likes to go ahead (yu, 豫) of the man, and then cannot wait for him and comes back to meet him, and goes back and forth like this the whole day long. This is how yu comes to mean not settled, and so it is called youyu. Some take the Erya, which says, "The you is like the ji (麂) and climbs trees well." You is the name of a beast, and as soon as it hears a man's voice it climbs a tree beforehand, and goes up and down like this, and so it is called youyu. The fox as a beast is also full of suspicion, and only dares to cross when it has listened under the ice of the Yellow River and heard no sound of running water. The common saying today is, "The fox doubts, the tiger divines." That is the sense.

The Zuo zhuan says, "The marquis of Qi had jie (痎), and it became shan (痁)." The Shuowen says, "Jie is the ague that comes on every second day. Shan is ague with fever." I have looked into it. The marquis of Qi's illness came on every other day at first and then grew heavier, and so it became the care of the feudal lords. In the north they still call it jienüe (痎瘧) today, and the sound is jie (皆). But the copies current in the world mostly write jie (痎) as jie (疥), scabies, and Du Zhengnan gives no explanation either. Xu Xianmin reads it as jie (介), and the vulgar scholars smooth it over and say, "When a man has scabies it makes him shiver with cold, and this turns into ague." This is guesswork. Scabies and ringworm are small ailments, hardly worth discussing; and how should a man with scabies turn to ague?

The Book of Documents says, "Like shadow and echo (惟影響)." The Zhou li says, "The gnomon measures the shadow, the morning shadow and the evening shadow." Mengzi says, "He draws the shadow and loses the form." Zhuangzi says, "The penumbra asks the shadow." Characters like these should all be the jing (景) of guangjing (光景). A shadow in the shade comes into being by light, and so it is called jing. The Huainanzi calls it jingzhu (景柱), and the Guangya says, "The gnomon pillar hangs a shadow." All of these are the same. It was only in the Jin, in Ge Hong's Ziyuan, that san (彡) was first added at the side to make ying (影), with the sound the fanqie of yu (於) and jing (景). But the world at once set about altering the Book of Documents, the Zhou li, Zhuangzi and Mengzi to follow Ge Hong's character, and that is very wrong.

Duke Tai's Liu tao has the heavenly array, the earthly array, the human array and the array of clouds and birds. The Analects say, "Duke Ling of Wei asked Confucius about arrays (陳)." The Zuo zhuan says, "He drew up the fish-scale array." The common copies mostly write the mound radical with the che (車) of chariots beside it. I have looked into it. All these battle arrays are written with the chen of Chen and Zheng (陳鄭). The sense of marching array is taken from setting out in order (陳列); this is borrowing under the six categories of writing. Neither the Cang Jie and the Erya nor the character books of recent times has a separate character; only Wang Xizhi's Xiaoxue zhang alone puts che beside the mound. Even though it passes current among the common sort, it is not right to go back and alter the Liu tao, the Analects and the Zuo zhuan.

The Odes say, "The yellow bird flies and settles on the guanmu (灌木)." The Commentary says, "Guanmu is trees in a clump." This is the text of the Erya, and so Li Xun's note says, "Trees growing in a clump are called guan." The last chapter of the Erya also says, "Trees growing in a cluster are guan." Zu (族) too is gathering in a clump. This is why the old copies of the Odes in Jiangnan all write the cong (叢) of congju (叢聚); but the old form of cong is like the character zui (最), and scholars of recent times therefore altered it to zui and explained, "The tallest and longest among trees." I have looked into it. None of the several schools of the Erya or of the explainers of the Odes says this. Only Zhou Xuzhi's note on the Mao Odes reads it as the fanqie of cu (徂) and hui (會), and Liu Changzong's note on the Odes reads it as the fanqie of zai (在) and gong (公), or again of zu (祖) and hui (會). All of these are forced, and lose the gloss of the Erya.

Ye (也) is a word for ending a sentence and for helping out a clause, and the books have it everywhere. The classics and commentaries of Hebei leave this character out throughout, and among them are places where it cannot be done without. Take "bo ye zhi shu (伯也執殳)," "yu lü ye yu (於旅也語)," "Hui ye lü kong (回也屢空)," "feng, feng ye, jiao ye (風，風也，教也)," and in the Commentary on the Odes "bu ji, ji ye; bu nuo, nuo ye (不戢，戢也；不儺，儺也)" and "bu duo, duo ye (不多，多也)." In cases like these, if this character is pared away the text is left crippled and short. The Odes say, "Green, green is your collar (靑靑子衿)." The Commentary says, "The green jin is the green collar, the dress of a student." I have looked into it. In old times the slanting collar ran down and joined the lapel, and so the collar was called jin (衿). Sun Yan and Guo Pu in their notes on the Erya, and Cao Dagu in the note on the Lienü zhuan, all say, "Jin is the crossed collar." Since the Ye copies of the Odes have no ye, the scholars there went wrong and said, "Qingjin and qingling are the names of two places on the garment, and both are trimmed with green," and used this to explain the two words "green, green." Their error is great. There are also vulgar students who, having heard that the classics and commentaries sometimes need the character ye, put it in wherever they please, and always in the wrong place, which is more laughable still.

There is a note by Shucai (蜀才) on the Yi, and the scholars of Jiangnan have ended by not knowing who this was. Wang Jian's Sibu mulu does not give the name, and puts down only, "A man after Wang Bi." Xie Jiong and Xiahou Gai had both read several thousand volumes, and both suspected it was Qiao Zhou. But the Shu shu of the Li house, also called the Han zhi shu, says, "The surname was Fan, the name Changsheng, and the man called himself Shucai." In the south, after the house of Jin crossed the river, all the records of the north were called forged books and were not thought worth reading, and so this was never seen.

The Wangzhi chapter of the Book of Rites says, "luo gu gong (臝股肱)." Zheng Xuan's note says, "It means rolling up the clothes and baring the arms and the shins." The books today all write the huan (擐) of putting on armour. Xiao Gai, erudite of the National Academy, said, "Huan should be written xuan (揎), with the sound xuan (宣). Huan is a word for putting a thing on through, not for baring the arm." Looking into the Zilin, Xiao's reading is right, and Xu Yuan's reading of it as huan (患) is wrong.

The Han shu says, "Tian Ken (田肯) offered congratulations to the emperor." The Jiangnan copies all have the character xiao (宵). Liu Xian of Pei had read widely in the classics and the records and was especially deep in Ban Gu's Han shu, and under the Liang was called the Sage of the Han shu. Liu Xian's son Liu Zhen did not let the family calling fall. In reading Ban Gu's history he called it Tian Ken. Emperor Yuan of the Liang once asked him about it, and he answered, "There is no sense to look for here; only that in the old copy of my house xiao has been changed to ken with orpiment." Emperor Yuan could not gainsay it. When I came north of the river, I found the copies had ken.

The appraisal of Wang Mang in the Han shu says, "A purple colour and a frog's croak, an odd remainder in an intercalary place (閏位)." What it means is that this was not a colour of black and yellow, and not a sound in tune with the pitch pipes. Lately there was a scholar of very high repute who went so far as to say, "Wang Mang not only had the shoulders of a kite and the glare of a tiger, but a purple face and the croak of a frog." This too is a mistake.

The ce of jiance (簡策) is written with ci (朿) under the bamboo radical, but in the clerical script of a late age it looks like the song (宋) of Qi and Song, and there are also copies that end by writing jia (夾) under the bamboo. It is the same as the character ci (刺), which should have ci (朿) at its side but is now also written with jia. Xu Xianmin's Chunqiu yin and Li yin end by taking jia (筴) for the correct character and ce (策) for the sound, which is upside down indeed. The Shiji again writes the character xi (悉) wrongly as shu (述), and the character du (妒) wrongly as gou (姤), and Pei Yin, Xu Guang and Zou Dansheng all read the character xi as shu and the character du as gou. If that is so, then one might as well take hai (亥) for the sound of the character shi (豕), and di (帝) for the sound of the character hu (虎).

Zhang Yi says, "Fu (虙) is the Fuxi of today." Meng Kang's note on the old text of the Han shu also says, "Fu (虙) is the present fu (伏)." But Huangfu Mi says, "Fuxi is sometimes called Fuxi (宓羲)." Looking through the classics, the histories and the apocrypha, there is no such title as Fuxi (宓羲) anywhere. The character fu (虙) follows hu (虍), and the character fu (宓) follows □〔去掉必〕 under the roof radical, and below both is bi (必). Copyists of a late age ended by writing fu (虙) wrongly as fu (宓), and the Diwang shiji thereupon set up a further name, that is all. How is this proved? Confucius's disciple Fu Zijian (虙子賤) was magistrate of Shanfu, and was of the line of Fuxi; in the common form the name too is written fu (宓), and sometimes the mountain radical is added. The city of Yongchang commandery in Yanzhou today is the old land of Shanfu, and at its east gate stands Zijian's stele, set up under the Han, which says, "Fu Sheng (伏生) of Jinan was of the line of Zijian." From this it is plain that fu (虙) and fu (伏) have been interchangeable characters from of old, and that writing fu (宓) is the mistake.

The Grand Historian's record says, "Better be a cock's mouth than an ox's hinder part." This is an abridgement of the Zhanguo ce. I have looked into it. Yan Du's Zhanguo ce yinyi says, "Shi (尸) is the master among the fowls. Cong (從) is the ox's calf." If so, kou (口) should be shi (尸), and hou (後) should be cong (從); the common copyists have got it wrong.

Ying Shao's Fengsu tong says, "The Grand Historian's record: 'Gao Jianli changed his name and surname and hired himself out as a servant, and hid himself at work in Songzi. After a long while of bitter labour he heard a guest striking the zhu in the hall of the house, and his skill itched in him (伎癢) and he could not keep from speaking out.'" I have looked into it. Jiyang means to carry one's skill within and to itch in the belly with it. That is why Pan Yue's Rhapsody on Shooting Pheasants also says, "In vain the heart is vexed and the skill itches." Today the Shiji everywhere has "pacing about (徘徊)," or "wandering to and fro, he could not keep from speaking out," and this is an error of the common transmitters and copyists.

The Grand Historian, discussing Ying Bu, says, "The rising of the disaster began with his favourite concubine, grew out of dumei (妒媚), and came at last to the ruin of his kingdom." The Waiqi zhuan of the Han shu also says, "He was brought in the end to be put to death for the dumei of a favoured concubine." Both these mei (媚) should be mao (媢); mao too means jealousy, and the sense is found in the Book of Rites and the Sancang. Moreover the Wuzong shijia also says, "The queen of King Xian of Changshan was dumao." Wang Chong's Lunheng says, "When a jealous husband and a jealous wife are born, then there is anger and quarrelling and going to law." From this it is all the clearer that mao is another name for jealousy. The root of Ying Bu's death was his suspicion of Ben He, so it cannot be called mei (媚).

The Basic Annals of the First Emperor in the Shiji say, "In the twenty-eighth year the chancellors Wei Lin (隗林) and Wang Wan and the rest deliberated by the sea." The copies all have the lin (林) of shanlin, forest. In the fifth month of the second year of Kaihuang a man of Chang'an digging in the ground found an iron steelyard weight of the time of Qin, with two inscriptions at its side cut into a copper coating. The one said, "In the twenty-sixth year the emperor wholly annexed the feudal lords of the empire and the common people were greatly at peace, and he set up the title of emperor, and thereupon ordered the chancellors Zhuang (狀) and Wan (綰) that whatever was doubtful in the standards of measure through not being uniform should all be made clear and made one." Forty characters in all. The other said, "In the first year an order was given to the chancellors Si (斯) and Quji (去疾): the standards of measure were all made by the First Emperor, and all carry a cut inscription. Now that the title has been taken over, if the cut inscription does not name the First Emperor, then after a long time it will look as though a later heir had made them, and the merit achieved and the fullness of virtue will go unnamed. Cut this edict on the □ left side, that there be no doubt." Fifty-eight characters in all, of which one is worn away and fifty-seven are to be seen, plain and clear. The writing is partly in the old clerical script. I was ordered to copy it out and read it, and sitting face to face with Li Delin, director of the palace secretariat, I saw this weight; it is now in the government store. The character in "the chancellor Zhuang" is the zhuang (狀) of zhuangmao, appearance, with quan (犬) beside the zhuang radical. From this it is plain that the "Wei Lin" of the common copies is wrong, and that it should be Wei Zhuang (隗狀).

The Han shu says, "Within and without there is zhifu (禔福)." The character should follow shi (示). Zhi means at peace, and its sound is the chi of chibi (匙匕), a spoon; the sense is found in the Cang Jie, the Erya and the Fangyan. The scholars of Hebei all say so. But the copies of Jiangnan mostly follow the hand radical by mistake, and those who write prose match it in couplets and all take it in the sense of tiqie (提挈), lifting and holding, which I fear is wrong.

Someone asked, "The note to the Han shu says, 'Because the father of the Empress Dowager Yuan was named Jin (禁), jinzhong (禁中) was called shengzhong (省中).' Why was sheng put in place of jin?" I answered, "I have looked into it. In the Zhou li, under the palace warden, 'He has charge of the warnings and orders and the jiujin (糾禁) of the royal palace.' Zheng Xuan's note says, 'Jiu is like cutting, and like examining.' Li Deng says, 'Sheng (省) is to examine.' Zhang Yi says, 'Sheng is the present shengcha (省□〔示改言〕).' If so, the two fanqie, xiaojing (小井) and suoling (所領), both take the meaning of examining. Since that place always had guards on watch examining it, sheng was put in place of jin. Cha (□〔示改言〕) is the old form of the character cha (察)."

The Annals of Emperor Ming of the Han say, "A school was set up for the lesser marquises of the four surnames (四姓小侯)." I have looked into it. When Emperor Huan was capped, he also gave silk to the four surnames and to the lesser marquises of the Liang and the Deng, from which it is plain that these were all consort kin. In Emperor Ming's time the consort kin were the Fan, the Guo, the Yin and the Ma, and these were the four surnames. They are called lesser marquises, perhaps because they received their fiefs while young in years, and so a school had to be set up for them. Or it is said that, attending the sacrifices and coming crowded to court, they were marquises but not full marquises, and so were called lesser marquises. The Book of Rites says, "The lesser lords of the several regions," which is the same sense.

The Hou Han shu says, "A stork carried three shan (鱔) fish." It is often borrowed as the zhan (鱣) of zhanwei (鱣鮪), and the vulgar scholars therefore call it the zhan fish. I have looked into it. Emperor Wu of Wei's Sishi shizhi says, "The zhan fish is as big as a five-peck box and a full fathom long." Guo Pu's note on the Erya says, "The zhan is two or three fathoms long." How should a stork be equal to one of them, let alone three? The zhan is moreover plain ash-grey, with no markings. The shan at its longest is not over three feet, and at its thickest not over three fingers, and is yellow with black markings. That is why the lecture-master said, "The snake and the shan are the emblem on the robes of ministers and grandees." The Xu Han shu and the Soushen ji also tell this story, and both write the character shan (鱔). Sun Qing says, "Fish, turtles, loaches and zhan." Han Fei and the Shuoyuan both say, "The zhan is like a snake, the silkworm like a caterpillar." All these write the character zhan (鱣). Borrowing zhan (鱣) for shan (鱔) is a thing of long standing.

The Hou Han shu says, "When the harsh official Fan Ye was governor of Tianshui commandery, Liangzhou made a song about him: 'Better to look on the den of a suckling tigress than to enter the offices of the Ji government.'" But in the books of Jiangnan the xue (穴), den, is everywhere written wrongly as liu (六), six. The scholars follow on and go astray without waking to it. That tigers and leopards live in dens is plain enough, which is why Ban Chao said, "If you do not grope into the tiger's den, how will you get the tiger's cub?" How should it be a question of six or seven?

The life of Yang You in the Hou Han shu says, "The wind blew the xiaofei (削肺)." This is the shaving (柿) pared from a wooden tablet. In old times, when writing went wrong, it was pared away, which is what the Zuo zhuan means by "he pared it off and threw it away." Or the tablet itself is called xiao; Wang Bao's Tongyue says, "He writes on xiao instead of on tablets," and Su Jing's letter says, "With the skill of old for rubbing and polishing, binding and paring." These are all the proof. The Odes say, "They fell the trees, huhu (滸滸)." The Mao Commentary says, "Huhu is the look of shavings." The historians borrowed for it the character of ganfei (肝肺), liver and lungs, and the common copies therefore all write the fu of fula (脯臘), dried meat, or the bu of fanbu (反哺). The scholars then explain, "Xiaobu is the name of a screen." There is no evidence for it, and it is nonsense besides. This is only divination by the angles of the wind. The Fengjiao shu says, "The wind of the common people brushes the ground, raises the dust and rolls the shavings." If it were a screen, how could it be rolled?

The Sanfu juelu says, "Fan Zhonggong, grandmaster of the front company, kept salt, fermented beans and suanguo (蒜果) all in one tube." Guo (果) should be written as the guo (顆) of Wei Guo (魏顆). In the northern country a lump of anything is generally called a guo, and suanguo (蒜顆), a clove of garlic, is ordinary speech among the common people. That is why the King of Chensi's Rhapsody on the Hawk and the Sparrow says, "Its head like a clove of garlic, its eye like a split peppercorn." And a Daoist book says, "With closed lips they chant the sutra, and the sound is small and fine; from their eyes the tears come out like beads, □〔石果〕." The characters differ, but the sound and the sense are much the same. In Jiangnan they only call it suanfu (蒜符) and do not know that it is called guo. The scholars have handed it down from one to another and read it as the guo (裹) of guojie (裹結), wrapping, and say that the salt and the garlic were wrapped in one bundle and put inside the tube. The Zhengshi xiaofan yinyi again reads suanguo as the fanqie of ku (苦) and ge (戈). All of these are wrong.

Someone came and asked me, "In Jiang Ji's memorial in the Wei zhi it says, 'the worn and □〔支力〕 people.' What character is this?" I answered, "To my mind □〔支力〕 is simply the □〔危皮〕 of □〔危皮〕juan (倦), weary. Zhang Yi and Lü Chen both say, 'With the dao of daojian (刀劍), sword, beside zhi (支), it is also the character ji (剞).' Whether Master Jiang made up for himself a form with the li (力) of jinli (筋力), strength, beside zhi, or borrowed the character ji, I do not know; but in the end it should be read as the fanqie of jiu (九) and wei (僞)."

The Jin zhongxing shu says, "Yang Man of Taishan was always slack and unrestrained and played the gallant, drank wine and kept no measure, and in Yanzhou they called him Tabo (濌伯)." Nowhere is there sound or gloss for this character. Emperor Xiaoyuan of the Liang often said to me, "I never knew it. Only Zhang Jianxian taught me, and reads it as the ta of tageng (嚃羹), gulping soup. Ever since I have followed that, though I do not know where it comes from either." Jianxian is the posthumous title of Zhang Zuan, governor of Xiangzhou, who was called a great scholar in Jiangnan. I have looked into it. He Fasheng's age was very close to the present, and he must have had it by word of mouth from the old men. There is also a word tata (濌濌) among the common people, meaning that there is nothing it does not extend and nothing it does not take in. Gu Yewang's Yupian wrongly writes it with ta (沓) beside the black radical. Gu was widely learned, but still stands below Jianxian and Xiaoyuan, and both of those said it was the chong (重) radical. In the several copies I have seen there is none that writes it with black. Chongta (重沓) has the sense of much and ample, piled up and thick; following black there is no sense in it at all.

In the words of the old yuefu song, three sons are spoken of first and then the three fu (婦); fu is a name used with respect to the husband's parents. Its last stanza says, "Sit at ease, zhangren; the tuning of the strings is not yet done." In old times a son's wife served her husband's parents and was at their side morning and evening, no different from their own children, and so these words were used. Zhangren is also a term for an elder, and in common speech today a grandfather and a father who have died are still called zhangren. It may be suspected, too, that zhang (丈) should be written da (大), for in the customs of the north a son's wife calls her husband's father daren gong (大人公). Zhang and da are easily confused. Men of letters in recent times have written a good many Three Wives poems, taking it in the sense of pairing the principal wife and all a man's other wives together, and adding to it the language of Zheng and Wei. How can gentlemen of the Daya go so far astray?

The old yuefu song of Baili Xi says, "Baili Xi, five ram skins. I think of the time we parted: we boiled the brooding hen and chui the door bar. Today you are rich and noble and have forgotten me!" Chui (吹) should be written as the chui (炊) of cooking. I have looked into it. Cai Yong's Yueling zhangju says, "The jian (鍵) is the male pin of the bolt, which stops the door leaf; some call it yanyi (剡移)." If so, they were poor at that time and used even the wood of the door bar for firewood to cook with. The Shenglei writes yan (扊), and some write dian (扂).

The Tongsu wen is inscribed in the world as "made by Fu Qian of Henan, styled Zishen." Fu Qian was a man of the Han, yet his preface cites Su Lin and Zhang Yi, and Su and Zhang were both men of Wei. Moreover, before Zheng Xuan no one understood fanqie at all, while the fanqie sounds of the Tongsu wen agree very closely with the common sounds of recent times. Ruan Xiaoxu again says it was "made by Li Qian." In Hebei every house keeps a copy of this book, and there is none that says Li Qian made it. Neither the Jin zhongjing bu nor the Qizhi has it in its catalogue, and in the end there is no knowing who composed it. Yet its wording and its sense are apt, and it is truly the work of high talent. Yin Zhongkan's Changyong zixun also cites the common explanations of Fu Qian, but that book no longer exists, so there is no knowing whether it is the Tongsu wen or something different, or whether there was yet another Fu Qian. It cannot be made clear.

Someone asked, "The Shanhai jing was recorded by Yu of the Xia and by Yi, yet it has Changsha, Lingling, Guiyang and Zhuji, and there are not a few commanderies and counties like these. How is this?" I answered, "The gaps in the histories are of long standing, and besides, the men of Qin destroyed learning and Dong Zhuo burned the books, and the confusion of the records does not stop here. To take a comparison: the Bencao was set out by Shennong, yet it has the names of commanderies and counties such as Yuzhang, Zhuyai, Zhaoguo, Changshan, Fenggao, Zhending, Linzi and Pingyi, given as the places the drugs come from. The Erya was made by the Duke of Zhou, yet it says, 'Zhang Zhong was filial and brotherly.' Zhongni edited the Chunqiu, yet the text records the death of Kong Qiu. The Shiben was written by Zuo Qiuming, yet it has King Xi of Yan and Gaozu of the Han. The Suoyu from the Ji tomb carries the stele of Qinwang. The Cangjie pian was made by Li Si, yet it says, 'Han joined the empire together and set the lands within the seas side by side; Xi and Qing were overthrown, and the rebels were struck down and destroyed.' The Liexian zhuan was made by Liu Xiang, yet its appraisal says that seventy-four of the men came out of the Buddhist sutras. The Lienü zhuan was also made by Liu Xiang, and his son Liu Xin further made the eulogies, which end with Queen Dao of Zhao; yet the text has Lady Han of Gengshi, Empress Ma of Mingde, and Lady Yi of the Liang. All these were foisted in by later men, and are not the original text."

Someone asked, "Why does the Donggong jiushi call the chiwei (鴟尾) ciwei (祠尾)?" I answered, "Zhang Chang was a man of Wu who did not much look into old things; he set down notes as it suited him, following the errors of local speech and making up characters, that is all. The men of Wu call cisi (祠祀), sacrifice, chisi (鴟祀), and so he put ci (祠) in place of the character chi (鴟). They call gan (紺) jin (禁), and so he wrote jin beside the silk radical in place of the character gan. They call zhan (盞) by the fanqie of zhu (竹) and jian (簡), and so he wrote zhan (展) beside the wood radical in place of the character zhan. They call the character huo (鑊) by the character huo (霍), and so he wrote huo beside the metal radical in place of the character huo. Again he wrote huan (患) beside metal for the character huan (鐶), gui (鬼) beside wood for the character kui (魁), shu (庶) beside fire for the character zhi (炙), and mao (毛) under ji (旣) for the character ji (髻); for jinhua (金花) he wrote hua (華) beside metal, and for chuangshan (窗扇) he wrote shan (扇) beside wood. Of this sort of wilfulness there is no little."

The questioner asked again, "What sort of thing is the liuse ji□〔糹畏〕 of the Donggong jiushi, and how should it be read?" I answered, "I have looked into it. The Shuowen says, 'Jun (莙) is the ox-weed, and is read like wei (威).' The Yinyin gives 'the fanqie of wu (塢) and gui (瑰).' This is what Lu Ji called 'the clustered weed, whose leaf is like mugwort.' Guo Pu's note on the Sancang also says, 'Yun (蘊) is a kind of water weed, with fine leaves growing thick and downy.' There is such a thing in the water today: each joint is a few inches long, and the fine down is like silk thread, curling round prettily, and the long ones have twenty or thirty joints, and it is still called jun. Again, five-coloured threads are cut into inch lengths and laid crosswise between the strands of a cord and twisted in, in the likeness of the jun plant, and used to ornament things, and this too is called jun. At that time they made this jun on a dark blue six-coloured felt to ornament a woven belt, and Zhang Chang therefore made the form with wei (畏) beside the silk radical; it should be written wei (隈)."

Northeast of Bairen city there is a lone hill which no old book records. Only Kan Yin's Shisan zhou zhi takes the great forest that Shun entered to be this very hill, and on its top there is still a shrine to Yao today. Among the common people some call it Mount Xuanwu and some Mount Xuwu, and no one knows where the name comes from. Li Mushu and his brother Li Jijie, and Li Puji, of the great families of Zhao commandery, were men of learning too, yet none of them could settle the name of this hill in their own district. I was once an aide in Zhaozhou, and together with Wang Shao of Taiyuan I read the stele inside the west gate of Bairen city. The stele was set up by the people of Bairen county for their magistrate Xu Zheng in the time of Emperor Huan of the Han, and its inscription said, "On the hill is Quanwu□〔女改山〕, where Wang Qiao became an immortal." Only then did it become clear that this is Mount Quanwu□〔女改山〕. For the character quan (巏) there is no source at all. The character wu□〔女改山〕, by the several character books, is the mao (旄) of maoqiu (旄丘); and one sound of the character mao in the Zilin is the fanqie of wang (亡) and fu (付), so that, going with the common name, it should be read Quanwu (權務). When I came to Ye and told Wei Shou of it, Wei Shou marvelled greatly. He happened to be composing the stele inscription for the Zhuangyan monastery in Zhaozhou, and so he wrote, "the essence of Quanwu." That is, he used this.

Someone asked, "Why is one night five geng (更)? What does geng mean?" I answered, "From the Han and the Wei onward they said jia night, yi night, bing night, ding night and wu night; and again they said drums, first drum, second drum, third drum, fourth drum, fifth drum; and again first geng, second geng, third geng, fourth geng, fifth geng, all of them taking five as the measure. The Rhapsody on the Western Capital also says, 'Guarded by the office of the strict watch.' The reason is this. Suppose the first month stands at yin: at evening the handle of the Dipper points to yin, and at dawn it points to wu, and from yin to wu it passes five stations in all. In the months of winter and summer the length is uneven, long or short, but the stations are wide apart, so that at the fullest it does not pass six and at the shortest it does not fall to four, and the advance and the retreat always lie between the five. Geng means to pass through, to go over, and so it is called the five geng."

The Erya says, "Zhu (朮) is shanji (山薊)." Guo Pu's note says, "The zhu of today is like the thistle and grows in the hills." I have looked into it. The leaf of the zhu is like the thistle in shape, yet men of letters in recent times have gone so far as to read ji as the jin (筋) of jinrou (筋肉), sinew and flesh, and use it as a match for digu (地骨); I fear this loses the sense.

Someone asked, "The common name for a puppet is Guo Tu (郭禿). Is there an old story behind it?" I answered, "The Fengsu tong says, 'All the Guo family avoid the word tu (禿), bald.' There must have been a man of a former age surnamed Guo who had the disease of baldness and was made fun of in jest, and so later men made an image of him and called it Guo Tu, just as Wenkang was made in the likeness of Yu Liang."

Someone asked, "Why is the aide who handles prison cases called changliu (長流)?" I answered, "The Diwang shiji says, 'When Emperor Shaohao died, his spirit came down on the hill of Changliu, and in the sacrifices he presides over autumn.' I have looked into it. In the Autumn Office of the Zhou li the minister of justice has charge of punishments, and the post of changliu was only the bandit-catching clerk of the Han and the Wei. From the Jin and the Song onward it first became an aide's post, belonging above to the minister of justice, and so the dwelling of the autumn god was taken for a fine name for it."

A guest put it to his host: "The classics of today you say are all wrong, and what the Shuowen says you say is all right. Then is Xu Shen better than Confucius?" The host clapped his hands and laughed aloud and answered, "Are the classics of today all in Confucius's own hand?" The guest said, "Is the Shuowen of today all in Xu Shen's own hand?" The host answered, "Xu Shen checked with the six categories of writing and threaded it through by radicals, so that it could not go wrong, and if it did go wrong one would know. Confucius kept the sense and did not discuss the characters. Even the earlier scholars could alter the characters to follow the sense; how much more so in copying and handing down. Things like 'stopping the spear is wu (武),' 'turning zheng (正) over is fa (乏),' 'a vessel with insects is gu (蠱),' and 'hai (亥) has two heads and six bodies' in the Zuo zhuan, later men naturally cannot alter at will; how then should one dare to use the Shuowen to judge them right or wrong? And I do not hold the Shuowen alone to be right either; where its citations from the classics and commentaries differ from the text as it now stands, I do not dare to follow it. Again, Xiangru's Fengshan shu says, 'He dao (導) to the kitchen the one stalk with six ears, and took for the victim the beast whose two horns rise from one root.' Here dao is glossed as choosing, and Emperor Guangwu's edict, 'Not only the labour of rearing beforehand and choosing out,' is the same. Yet the Shuowen says, 'Dao is the name of a grain,' and cites the Fengshan shu as its evidence. There is no objection to a grain named dao existing on its own, but it is not what Xiangru used. 'A grain of one stalk and six ears to the kitchen' — how is that a sentence? Even if Xiangru's natural talent had been coarse and clumsy and he had forced out such a phrase, the line below would then have to read 'the unicorn, the beast whose two horns rise from one root,' and could not say 'victim.' I have laughed in my time at Xu as a mere plain Confucian who does not reach the body of good writing; that sort of thing is not to be relied on. But on the whole I bow to his book, for it is trimmed to a settled rule and cuts things open to the very root, and Zheng Xuan in annotating books often cites it for evidence. If one does not believe what it says, then all is dark, and one does not know what meaning there is in a single dot or a single stroke."

Those in the world who study the lesser learning do not have a grasp of both old and new, and insist on relying on the small seal script to set writings right. But could the Erya, the Sancang and the Shuowen have caught the whole of Cang Jie's original intent? They too added and took away with the ages, and differ from one another as much as they agree. How can the character books of before the Western Jin all be called wrong? Only let the rule of the forms be carried through, and let no one act at his own pleasure. In weighing right and wrong one must look very closely. Take "Zhongni ju (仲尼居)": of the three characters two are not in the correct form, for the Sancang adds qiu (丘) beside ni and the Shuowen puts ji (幾) under shi (尸). How can one follow that sort of thing? In old times the doubled characters did not exist and borrowing was common: zhong (中) was used for zhong (仲), yue (説) for yue (悅), shao (召) for shao (邵), xian (閑) for xian (閒). This sort needs no correcting. But there are outright errors which have become mere vulgarity: the side of luan (亂) made she (舌); yi (揖) with no er (耳) beneath; yuan (黿) and tuo (鼉) following gui (龜); fen (奮) and duo (奪) following guan (雚); dai (帶) put inside xi (席); xi (西) set on top of e (惡); pi (皮) placed outside gu (鼓); hui (毁) growing at the head of zao (鑿); li (離) paired with yu (禹); he (壑) written with huo (豁); wu (巫) mixed with the side of jing (經); gao (皋) split off from a piece of ze (澤); lie (獵) turned into ge (獦); chong (寵) changed into □〔上穴下龍〕; pian (片) added to the left of ye (業); qi (器) set under ling (靈); the character shuai (率), which has the sound lü (律) of its own, forced into a separate character; the character dan (單), which has the sound shan (善) of its own, split at once into a different character. This sort of thing cannot be left untreated. When I first looked at the Shuowen long ago, I laughed at the characters of the day; but to follow the correct forms was to fear that no one would know them, and to follow the common forms went against my sense of what was wrong, so that I could hardly set brush to paper. As what I saw grew wider I came to know how to move with change, and I have mended my old obstinacy by about half. In composing literary work I still choose the forms that come somewhat near; in office papers and in the letters of common life it is as well not to go against custom.

I have looked into it. The character geng (亙) of migeng (彌亙) follows zhou (舟) between two er (二), as in the Odes, "geng zhi juju (亙之秬秠)." The clerical script of today has turned zhou into ri (日); and He Fasheng's Zhongxing shu took the zhou between the two er for the character of zhouhang (舟航), which is wrong. The Chunqiu shuo takes ren (人), shi (十), si (四) and xin (心) for de (德); the Shi shuo takes er (二) under tian (天) for you (酉); the Han shu takes huoquan (貨泉) for baishui zhenren (白水眞人); the Xinlun takes jinkun (金昆) for yin (銀); the Guozhi takes kou (口) above tian (天) for Wu (呉); the Jin shu takes huangtou xiaoren (黃頭小人) for gong (恭); the Song shu takes zhaodao (召刀) for shao (邵); the Cantong qi takes ren (人) carrying gao (告) for zao (造). Examples of this kind are the wrong-headed talk of the arts of number, borrowing and fastening things together with jest and laughter mixed in. It is like turning the character gong (貢) into xiang (項), or making chi (叱) into bi (匕). How can the sound and reading of characters be settled by such means? The acrostic poems and rhapsodies of Pan Yue, Lu Ji and the rest, the books of divining boards and of splitting characters, and Bao Zhao's riddle characters, are all got up to suit the taste of the day, and are not enough to be argued over as form and sound.

Xing Fang of Hejian said to me, "The life of Jia Yi says, 'At midday one must □〔上彗下火〕,' and the note says, '□〔上彗下火〕 is pu (暴).' I once saw a man explain it thus: 'This is the sense of sudden and violent; it says only that midday does not stay a moment but tilts over all at once.' Is this explanation right?" I said to Xing, "These words come originally from Duke Tai's Liu tao. If one looks into the character books, in old times the pu of pushai (暴曬), drying in the sun, and the pu of puji (暴疾), sudden illness, were alike, and differed only a little below, and later men added the ri (日) at the side of their own accord. It says that at midday one must dry things in the sun, and that if not, the time is lost. Jin Zhuo has already explained it in full." Xing Fang laughed and gave way and withdrew.