The work known as the Analects is mainly a collection of sayings and conversations of Confucius. In the time of Emperor Wu (140-87 BC) of the Han there were three versions of the work, the Lulun, the Qilun and the Gulun. In the first century BC, Zhang Yu taught a version known as Lunyu according to Marquis Zhang, which incorporated readings from both the Lulun and the Qilun. Zheng Xuan (AD 127-200) further adopted readings from the Gulun. The text that has come down to us is that of He Yan (AD 190-249) in his Lunyu jijie (Collected Commentaries on the Lunyu). In the Analects of Confucius, three texts are commonly appealed to as sources of authoritative utterance: the Shijing, the Shujing and the Yijing (Book of Changes). While the Analects thus contain a distinct notion of authoritative texts, there is no reference to an explicit classical canon of any kind (Waley: 12).
The Analects: Texts and Commentaries
The work known as the Analects is mainly a collection of sayings and conversations of Confucius. In the time of Emperor Wu (140-87 BC) of the Han there were three versions of the work, the Lulun, the Qilun and the Gulun. In the first century BC, Zhang Yu taught a version known as Lunyu according to Marquis Zhang, which incorporated readings from both the Lulun and the Qilun. Zheng Xuan (AD 127-200) further adopted readings from the Gulun. The text that has come down to us is that of He Yan (AD 190-249) in his Lunyu jijie (Collected Commentaries on the Lunyu).
The extant Analects is in twenty books. Of the three early versions, only the Lulun had twenty books; the Qilun and the Gulun both had twenty-two books, though the extra books were not identical. According to the Xinlun of Huan Tan (24 BC-AD 56) the order of the chapters in the Gulun was different and there were more than four hundred variant readings. Lu Deming (AD 556-637) of the Tang dynasty also remarked that in the Qilun, besides the two extra books, 'the chapters and verses were considerably more numerous than in the Lulun'. Some of the variant readings were recorded by scholars before the three versions were lost and these have been collected by textual critics over the centuries, but these consist mainly of variant forms of graphs. Only a handful affects the interpretation of the text (Legge: 63).