Learn · Essay · 12 min read · Updated September 2026

The texts

What the tradition actually wrote down, and when. The burial classic and its three sections, the house-siting classics and the canon that kept one of them, the manuals of the compass schools, the shelf the imperial librarians put them all on and what they wrote beside it, the Western accounts, and the paperbacks, with what each is good for and what it cannot be asked to say.

Oriented SW 237°

Earth

Fig. 01 Plaster, weathered. The tradition’s oldest text is a few pages long and has been read for seventeen centuries.

Feng shui is usually met second-hand: a rule, a corner, a thing a consultant said. Behind the rules is a small shelf of texts, some of them very old, most of them short, and none of them saying quite what the rules attributed to them say. This essay walks the shelf in order, from the fourth-century burial classic to the paperbacks of the 1990s, says for each what it is, who wrote it as far as anyone knows, and what it can and cannot be made to bear, and then shows where the imperial librarians of the eighteenth century put the whole shelf, which tells you what the tradition’s own scholars thought of it.

The shelf, in order

Fig. 01 The texts, on one axis. Above the line, the Chinese classics and manuals, from the Zangshu to the Flying Star treatise of the 1920s; below it, the Western accounts and the modern books through which most readers now meet the tradition. The gap between the two rows is the gap between what the tradition wrote and what has been written about it. The axis is stretched after 1500 and again after 1850, where most of the texts are. Dates are approximate where the texts are: the Zangshu’s attribution to Guo Pu is recorded much later than his lifetime, and the Huangdi Zhaijing is older than its earliest surviving copy. The dot marks the century, not the year.

The burial classic

Tradition · The Zangshu

The Book of Burial, 葬書, is the text the tradition itself names first: about two thousand characters, a few pages in translation, in three sections, the inner, the outer and the miscellaneous, on the siting of graves. It opens with the sentence that gives feng shui its name, that qi rides the wind and scatters and stops when it meets water, and it sets out the four animals, the dragon veins, the lair where qi gathers and the reading of hills and watercourses that the Form School still uses. It is attributed to Guo Pu, a scholar, poet and diviner of the Jin dynasty who died in 324, executed for advising against a rebellion; the attribution is first recorded centuries later, in the Song, and the text as we have it may be a Tang or Song compilation of older material with the master’s name attached. Paton’s translation, with the commentaries, is the one to read, and the yin house page draws the valley the text describes.

Everything on this site about the shape of a site descends from the Zangshu, and almost nothing on this site about a room does. The text is about the dead. Its transfer to the houses of the living, the yang zhai, was the work of later writers who kept its vocabulary and changed its object, and the transfer is where the tradition’s rules for rooms begin.

The house-siting classics

The Huangdi Zhaijing, 黃帝宅經, the Yellow Emperor’s Classic of the Dwelling, is the earliest sustained treatment of houses rather than graves. Its name is a convention, since the Yellow Emperor is the name every early technical text borrowed; its date is Tang or earlier; and its copy in the Daoist canon, the great collection of Daoist scripture printed in 1445, is the earliest that survives, which is why a book about houses is found among the liturgies of a religion. It divides houses into yin and yang by their orientation, twenty-four each, and reads their fortunes by the year, and it is the ancestor of every later manual that tells a household what to do with a house it already has.

Yang Yunsong, a court official of the late Tang who is said to have carried the imperial siting texts south when the capital fell to Huang Chao’s rebellion in 880, is the Form School’s named founder, and the Han Long Jing, 撼龍經, the Shaking Dragon Classic, and its companion the Yi Long Jing, the Doubtful Dragon Classic, are the texts attributed to him: long verse treatises on reading mountains as dragons, their veins, their joints and their lairs, with the nine star-shapes of peaks that the five elements essay draws. They are landform manuals, and the landform reading on the outside page is their reading at the scale of a garden.

The compass manuals

The Compass schools have their own shelf, and it is later. The Qingnang Jing, 青囊經, the Classic of the Blue Bag, is the short foundational text the San He tradition attributes to Huang Shigong, a figure of Han legend; the attribution is a way of claiming antiquity, and the text is read through Yang Yunsong’s commentary, the Qingnang Aoyu. The Bazhai Mingjing, 八宅明鏡, the Bright Mirror of the Eight Mansions, is a Qing text of the eighteenth century and the source of the Eight Mansions method as it is now taught: the kua number, the East and West groups, the eight directions for each. The Dili Wujue, 地理五訣, the Five Secrets of Geography, published by Zhao Jiufeng in 1786, is the manual through which the San He school’s use of the luopan became standard, and the luopan page takes its rings apart. And the Shen Shi Xuankong Xue, 沈氏玄空學, Master Shen’s Study of the Mysterious Void, compiled from the notes of Shen Zhureng and published by his son in 1927, is the text that made Flying Stars the dominant method of the twentieth century.

The shelf in the imperial library

Fig. 02 Where the siting books sat in the imperial library. The Complete Library of the Four Treasuries, compiled for the Qianlong emperor between 1773 and 1782, sorted every work it kept into four divisions: the Classics, the Histories, the Masters and the Collections. The Masters division held fourteen categories, among them the Confucians, the Legalists and medicine, and among them the numerical arts, which held seven kinds: astrology, the almanac, the theory of yin, yang and the five phases, divination, the books of fate and physiognomy, and, its own kind, the reading of houses and graves. That is the shelf the Zangshu, the Zhaijing, Yang Yunsong’s dragon classics and the Ming books on the yang house sat on, between the astrologers and the fortune-tellers, and the librarians’ notes on them are among the most sceptical things ever written about the practice by people who had read all of it. The catalogue’s notes on the numerical arts say that the arts are old, that their books are late, and that the compass school in particular had multiplied its rules past checking; Smith’s survey of the divinatory arts takes its map of the field from the same catalogue. The tree omits the categories on either side of the ones drawn.

The tradition’s own scholars catalogued the shelf, and where they put it is worth knowing. The Complete Library of the Four Treasuries, compiled for the Qianlong emperor between 1773 and 1782 by several hundred scholars from the books of the whole empire, sorted everything it kept into four divisions, the Classics, the Histories, the Masters and the Collections, and within the Masters into fourteen categories. The siting books went into the category of the numerical arts, beside astrology, the almanac, divination and physiognomy, in a kind of their own called the reading of houses and graves, and the editors wrote a note on each. Smith’s survey of the divinatory arts takes its map of the field from the catalogue, and the notes are the most level thing ever written about the practice by people who had read all of it: the arts are old, the books are late, the attributions to Guo Pu and Yang Yunsong cannot be verified, and the compass school in particular had multiplied its rules past anyone’s ability to check. The site’s own labels, tradition, school, interpretation, are a version of the librarians’ note.

The Western accounts

Ernest Eitel, a missionary in Hong Kong, published Feng-Shui: or, The Rudiments of Natural Science in China in 1873, the first sustained account in English, and gave the tradition the framing it has had in the West ever since: a would-be science, ingenious and wrong. J. J. M. de Groot’s volumes on The Religious System of China, from 1897, are longer, better informed and equally certain that the practice is superstition; they remain the best description of feng shui as it was practised in the nineteenth century, written by a man who watched it. Joseph Needham, in the second volume of Science and Civilisation in China in 1956, placed the tradition inside Chinese correlative thinking and gave it, for the first time in English, the respect of a history. Feuchtwang’s anthropological analysis of 1974 and Bruun’s fieldwork and survey of 2003 and 2008 are the modern scholarship the site leans on, and Smith’s histories of divination and of the Changes give the shelf its context.

The paperbacks

Feng shui reached the general reader in English in the 1980s, and the Chinese texts reached the Chinese reader again at the same time: the mainland reprinted the classics from the late 1980s after forty years in which they could not be sold, and the Taiwan and Hong Kong presses had kept them in print throughout. Stephen Skinner’s Living Earth Manual of 1982 was the first serious introduction to the classical system in English; Sarah Rossbach’s books of 1983 and 1987 carried Lin Yun’s Black Sect teaching into print and, with it, the door-aligned bagua and the life areas; Eva Wong’s manual of 1996 set out the Form and Compass methods with unusual clarity about which was which; and Lillian Too’s illustrated guide of the same year became the best-selling feng shui book in the world and the source of most of what the decorating press repeats. The site cites all four for what the modern schools say, and none of them as an authority on what the classical texts say.

The shelf sorted by subject rather than by date shows the same thing from the other side: how narrow each book is, and how late the ones about rooms are.

Fig. 03 The texts by subject. Fifteen books against seven subjects: a red dot is what a book is about, a ring is what it touches. Down the columns: the graves and the landform belong to the oldest texts; the house has its own classics from the Tang and the Ming; the compass and the calendar are Qing and later; the rooms and their cures are on the newest books; and the accounts of the practice are all Western. Across the rows: no text covers the subject, and the ones that come closest are the modern manuals. The subjects are the site’s reading of what each text is for, not a table of contents. Skinner and Rossbach are separated here, as they are not on the axis above: the first describes the classical system, the second carried the Black Sect method and its cures into English.

How the site uses them

The sources page lists every work cited and says what kind of authority each carries. The classics are cited for what the tradition holds; the compass manuals for what a school teaches; the Western scholars for the history and the practice; the paperbacks for the popular tradition; and the research, which is on a different shelf altogether, for what has been measured. The methodology page explains the six labels that keep them apart on the page.