Kǎogǔ tú 考古圖
An Illustrated Investigation of Antiquities by 呂大臨 (Lǚ Dàlín, 1027–1097, 宋, zhuàn 撰)
About the work
The single most influential Northern-Sòng jīnshí (epigraphy/antiquarian) work, in 10 juàn, by Lǚ Dàlín 呂大臨 (zì Yǔshū 與叔, of Lántián 藍田 in Shǎnxī, 1027–1097), the great Northern-Sòng Lǐxué scholar who first studied under Zhāng Zài 張載 (the Héngqú master) and then under Chéng Yí 程頤. The work was completed in Yuányòu 7 (1092, rénshēn 壬申), per the Kǎogǔ tú jì (the work’s zìxù preserved at the head). It systematically registers ancient ritual bronzes — zūnyídǐngduīyǒu type vessels — from a large number of Northern Sòng collections, both imperial and private, illustrating each piece, transcribing the inscription in kuǎnshí tracing, and giving Lǚ’s interpretive notes. The transmission has three layers: (1) the original Kǎogǔ tú in 10 juàn (1092); (2) the Xùkǎogǔ tú 續考古圖 in 5 juàn — the Sìkù editors establish this is a Southern Sòng continuation by an unnamed hand, since it cites Lǚ Dàlín in the third person and contains pieces dated Shàoxīng rénwǔ 壬午 (1162) and after — fully 70 years after Lǚ’s death; (3) the Shìwén 釋文 in 1 juàn, a phonological/script-style index to the inscriptions of the main 10 juàn. The WYG text used by the Sìkù editors derives from a Northern-Sòng cut version held by the Wúxí 無錫 Gù Chén 顧宸 family, later passing to Jì Zhènyí 季振宜 of Tàixīng, then to Xú Qiánxué 徐乾學 of Kūnshān, and finally borrowed by Qián Céng 錢曾 for transcription with a hand-copied illustration cycle (recorded in his Dúshū mǐnqiú jì 讀書敏求記). This WYG text contains a number of extra entries (a Kǒng Wénfù yǐndǐng in juàn 1; a yī-vessel in juàn 3; extra illustration and gloss material in juàn 4, 6, 8, 9, 10) — and its arrangement differs from the popularly-circulating texts. The work was the model for Wáng Fǔ’s KR3h0087 Xuānhé bógǔ tú of the same generation, which was assembled at the imperial court rather than by a private scholar.
Tiyao
We have respectfully examined: Kǎogǔ tú in ten juàn; Xùkǎogǔ tú in five juàn; Shìwén in one juàn — by Lǚ Dàlín of the Sòng. Dàlín, zì Yǔshū, of Lántián, in the Yuányòu (1086–1093) period a Mìshūshěng zhèngzì; his shìjì is attached to the Sòngshǐ biography of Lǚ Dàfáng. Chén Zhènsūn’s Shūlù jiětí records “Lǚ Dàlín’s Kǎogǔ tú in 10 juàn”; Qián Céng’s Dúshū mǐnqiú jì further records “beyond the 10 juàn there is also a Xùkǎogǔ tú in 5 juàn and Shìwén in 1 juàn, a Northern-Sòng cut-block obtained from the Wúxí Gù Chén family, later passing to Tàixīng Jì Zhènyí, then to Kūnshān Xú Qiánxué; Céng borrowed from Qiánxué to transcribe; the illustrations also commissioned from skilled hands without losing a háofā; paper and ink even better than the original print.” This text on collation against the impression marks is precisely Céng’s hand-transcribed copy. Comparing it with the popular texts of the day: juàn 1 has one extra Kǒng Wénfù yǐndǐng with 14 inscription characters and 51 gloss characters; juàn 3 has one extra gài (lid) illustration on the yīduī; juàn 4 has the Kāifēng Liú-family xiǎofānghú placed where the Mìgé fāngwén fānghú should be, and vice versa — the present popular texts have these transposed. Juàn 6 catalog adds a heading-line “pán, yí, yú, nǔ, gē, xuē” of one line. Juàn 8 has three extra Jade-Rope-Sword illustration with 115 shuō characters, and also extra Báiyù Yúngōu, Yùhuán, Yùjué illustrations one each. Juàn 9 has one extra Jīngzhào Tián-family lùlú dèng illustration with 47 shuō characters; the Rhinoceros Lamp Second illustration is significantly different from the popular text; the inner-stored huáněr zī has an extra gài illustration. Juàn 10 has the Xīnpíng Zhāngshì liánhuán dǐng illustration with no notes on “right place from where obtained” or measurements or inscription identification — all of which are lost and unverifiable; only the form remains here — 20 characters; further an extra Lújiāng Lǐshì jiāodòu illustration; the shòulú Second Illustration has 35 additional notes; the juàn-tail has an extra Qióngzhōu Tiānníng Temple monk-holding-mandate-amulet illustration two with 46 shuō characters. The juàn-head Dàlín zìxù, originally titled Hòujì, is appended at the juàn-tail. The other character-and-phrase variants we cannot enumerate exhaustively, but cross-checking the meaning of the text, all favour this edition.
The xùtú: juàn 1 has 20 vessels; juàn 2 has 22; juàn 3 has 26; juàn 4 has 20; juàn 5 has 12 — not arranged before-and-after by category — clearly the entries were added “as seen, recorded”; that is why the fifth juàn alone has so few entries. Some have inscriptions but no inscription-tracing; some have inscriptions but no reading-gloss. Their collectors’ names are recorded at the head of the illustration-gloss as “so-and-so acquired” — slightly different from the previous-illustrations practice of noting xìngmíng (collector-name) under the rubric. The Shìwén one juàn has Dàlín’s heading-words at the start: it takes the inscription characters, arranges them in the Guǎngyùn four-tone order; where there are variants, each is given a separate gloss; doubtful characters, xiàngxíng characters, characters whose radical is uncertain — these are appended at the end of the juàn. Dàlín’s Kǎogǔ tú was completed in Yuányòu rénshēn (1092), preceding the Xuānhé bógǔ tú — and its protocol is rigorous: where doubtful, it leaves blank — unlike the Bógǔ tú’s tendency to assert ancient attributions which become errors when checked. The yīduī entry, for example, is cited by Hú Ānguó’s commentary on the Chūnqiū “ChéngZhōu Xuānxiè huǒ” entry as gǔjīng (canonical-explanation) reference — proving its accounts trustworthy. Wú Qiūyǎn’s Xuégǔ biān states that this tú has two versions, in hēizì (black) and báizì (white) types (referring to inscribed-style differences); the hēizì version has a Yùntú at the end and lacks the Ményù zhì (jade rope-sword fitting); the báizì version has on the Bóshānlú an illustration of human-hands holding a chicken. The present text’s inscription is in the báizì style, yet the Bóshānlú illustration has neither human-hands nor chicken. Its Shìwén one juàn arranges characters in rhyme-order — that is Wú’s “Yùntú” — but juàn 8 actually has Lújiāng Lǐ-family Ményù zhì — so Wú’s version was less complete than this. Qián Céng calls it a piāonáng (silken-wrapper) rarity — not idle praise.
Only — the Xùtú five juàn are not in Chén’s catalogue, nor mentioned by Wú Qiūyǎn. Its second juàn cites “Lǚ Yǔshū says” and “the Kǎogǔ tú says”; the third juàn contains pieces acquired in Shàoxīng rénwǔ (1162) and after — therefore the book is after Shàoxīng 32 (1162) — too far apart from Dàlín’s lifetime. Likely a Southern-Sòng person continued Dàlín’s book but lost their name; Qián Céng concluded it was all by Dàlín — clearly without sufficient inspection. The Shìwén’s entries are all from the first ten juàn; the character explanations (e.g. xiè, xī) mostly agree with the túshuō. Only the Hé character — túshuō reads as Zhāng (matching Ōuyáng Xiū’s Jígǔ lù), but the Shìwén uses quēyí — a slight conflict; perhaps Dàlín had not finished revising. As to the tící’s saying that “ancient vessel inscriptions are not only different from xiǎozhuàn but also that the same character on the same vessel may have variant strokes or different radical positions” — citing the bǎi, bǎo, qí characters of the Bóbǎifù dùn, the guǐ character of the ShūGāofù guǐ, the zuò character of the Jìndǐng; and different vessels’ yí, zūn, shòu, wàn etc. — vessel-and-container strokes all have small variations: this is to know that ancient writing is not necessarily tóngwén (uniform script); only at the Qín did it standardise — so it is not all subsumable under xiǎozhuàn: also a tōnglùn (universal argument). Respectfully collated, Qiánlóng 46 (1781), tenth month.
Abstract
The Kǎogǔ tú (1092) is the founding work of Chinese jīnshí xué — the modern empirical study of bronze and stone antiquities — and the most influential single Northern-Sòng antiquarian text. Lǚ Dàlín, the great Lǐxué scholar of Lántián, brought kǎozhèng rigour to the project: the principle of yǒu yí zé quē 有疑則闕 (“where doubtful, leave blank”) distinguishes the Kǎogǔ tú from later compilations like the imperial KR3h0087 Xuānhé bógǔ tú. Lǚ’s self-preface, the Kǎogǔ tú jì of 1092, is one of the most important early statements of the antiquarian rationale in the Confucian intellectual tradition: he uses Zhuāngzǐ’s “chúgǒu” (straw-dog) and “lúnbiǎn” (wheelwright) parables to defend antiquarianism against the charge of mere curio-collecting, arguing that “observing the vessel and reading its inscription, in imitation we approach the yífēng of the Three Dynasties — yǐ yì nì zhì, perhaps exploring the principle of its making — to supplement the jīngzhuàn’s missing parts and correct the zhūrú’s errors.” The Xùkǎogǔ tú (Southern Sòng) and Shìwén (Lǚ’s own phonological index) extend the work into a complete jīnshí xué methodology. The Hong-Sòng-Wújiā-WYG transmission documented by the Sìkù editors gives this WYG a unique authority. The Sìkù editors also identify the Xùkǎogǔ tú as a separate Southern-Sòng work — a useful piece of kǎozhèng in itself.
Translations and research
- Sena, Yunchiahn C. Bronze and Stone: The Cult of Antiquity in Song Dynasty China. Seattle: University of Washington Press, 2019. [The standard Western-language study of the Kǎo-gǔ tú and the founding moment of Sòng antiquarianism.]
- Hsu, Ya-hwei 許雅惠. “Antiquaries and Politics: Antiquarian Culture of the Northern Song.” (Hsiang Lectures, 2014.)
- Ebrey, Patricia. Accumulating Culture: The Collections of Emperor Huizong. Seattle: University of Washington Press, 2008. [Sets the Kǎo-gǔ tú in its imperial-collection context.]
- Wáng Guówéi 王國維. Sòng-dài jīn-wén zhù-lù-biǎo.
- Sturman, Peter C. Mi Fu: Style and the Art of Calligraphy in Northern Song China. New Haven: Yale University Press, 1997.
- Wilkinson, Chinese History: A New Manual (under “Bronzes, Jades, and Antiquaries”) sets the Kǎo-gǔ tú as the typical starting-point for any modern study of bronze-vessel typology and pre-imperial inscriptions.
Other points of interest
The Sìkù editors’ detailed comparison of variant readings in this WYG against the popular Sòng/Yuán/Míng circulating texts is itself one of the most important pieces of kǎozhèng on the Sòng antiquarian tradition; the present WYG transmission, through the Wúxí Gù → Tàixīng Jì → Kūnshān Xú → Qián Céng chain, preserves a Northern-Sòng cut-block recension. The work’s identification of vessel-types and dating-by-inscription is the foundation for the entire later Chinese tradition of bronze-vessel typology (Lǚ → Wáng Fǔ → Xuē Shàngǒng → Wáng Qiú → Wú Dàchéng → Luó Zhènyù → Rong Geng).