Learn · The tradition · 15 min read · Updated September 2026

The yin house

Feng shui was a way of siting graves before it was a way of arranging rooms. The burial texts and the site they describe; the five things a geomancer read, in order; the belief that a good grave feeds the living; the lineages that fought over hills and the officials who tried to stop them; the tomb built as a valley in miniature; and what the tradition carried across from the dead to the living when it turned to houses.

Oriented N 015°

Water

Fig. 01 Green hills under a moving sky. The tradition read a grave by the hills round it before it read a house by anything.

The oldest text of the tradition is called the Book of Burial, and it is not a metaphor. For most of its history feng shui was a practice for choosing where to put the dead, and the vocabulary this site uses for a bedroom, the mountain behind, the arms at the sides, the open ground in front, the water that holds qi, was worked out at the scale of a valley for a grave. This page is about the yin house, the dwelling of the dead: what the texts say about it, how a geomancer read a site for one, what people believed a good grave did, who fought over them and who tried to stop it, and what crossed over when the tradition turned to the houses of the living.

The text is about graves

The Zangshu opens with the sentence that names the practice: qi rides the wind and scatters, and is held when it meets water; the ancients gathered it so that it would not scatter and guided it so that it would stop, and this they called feng shui. Everything that follows is instruction for a burial. The text describes the ridges along which qi travels, the dragon veins; the spot where they come to rest, the lair; the four animals that must stand about it, the azure dragon on the left, the white tiger on the right, the vermilion bird in front and the black tortoise behind; and the water that must embrace the site and leave it reluctantly. It says nothing about a door, a bed or a room. The attribution to Guo Pu, a scholar and poet of the fourth century, is recorded centuries after his death, and the text as we have it took its present shape later; but its subject was never in doubt. Paton’s translation, with the commentaries that grew around it, is the one to read.

Fig. 01 The classical site, as the burial texts draw it. Behind the grave, the ancestor mountain and, nearer, the parent mountain; between them and the lair run the dragon veins, the ridges the Form School reads as the channels of qi. The lair is the spot where the veins come to rest, and the grave sits in it, drawn as the horseshoe mound of a southern Chinese tomb. Two arms reach forward from the parent mountain, the dragon on the site’s left as it looks out and the tiger on its right, and close the bright hall, the open ground in front. Beyond it a low table hill, and far off the facing mountain. The water enters from the dragon side, curves round the front of the site and leaves by the water mouth, the narrow gap between two hills at the lower right. Every term in this drawing is a burial term first: the Zangshu and the Han Long Jing are about this site, not about a house. The tradition’s transfer of the same drawing to the living, mountain behind and water in front, seat in the lair, door as the mouth, is the subject of the article. North is up and the site faces south, as the texts assume; the drawing is a diagram of relations, not a survey.

How a site was read

The Compass school’s most widely used manual, the Dili Wujue of 1786, takes its title from the five things a geomancer must settle, and they are the Form School’s order of work as much as the Compass School’s: the dragon, the lair, the sand, the water and the facing. Read as a procedure, they describe a walk.

The dragon. Find the range the site descends from and follow its ridge down: is the vein alive, turning and rising and falling as it comes, or dead, a straight bare spur? The Han Long Jing classifies mountain forms under nine star-names, from the Greedy Wolf, a clean pointed peak, to the Broken Army, a ragged one, and reads the sequence of forms along a ridge as the dragon’s character.
The lair. Where the vein ends, look for the spot where it gathers: a slight rise on a slope, a shoulder with ground falling away on three sides, a place that the arms embrace. The lair is a point, not an area; the texts compare it to the point a physician’s needle finds on a body, and use the same word.
The sand. The sand is everything that stands around the lair: the arms, the table hill, the facing mountain, the nearer and further ridges. Each is judged by whether it shelters the lair or threatens it, whether it faces in or turns away, whether it is whole or broken.
The water. Where does it come from, how does it cross the front, where does it leave? The texts want it to arrive from the side, embrace the bright hall and go out by a narrow mouth. A stream that runs straight at the site, or straight away from it, is read as a fault, and the same fault indoors is a corridor.
The facing. Last, the compass: the direction the grave will face, read on the luopan against the twenty-four mountains, and reconciled with the dragon’s direction and the water’s exit by the school’s tables. The facing also fixed the date: a burial was timed by the almanac to a day and an hour that suited the site, and the geomancer chose both.

The order matters, and it is the order this site keeps for a house: landform first, the compass last, and never the compass instead of the landform. A geomancer who took a bearing before walking the ridge was, in the texts’ own terms, a fraud.

What a good grave was for

The belief that made the practice matter was resonance, ganying, the classical idea that things of the same kind act on one another at a distance. A body buried where qi gathered was thought to receive that qi in its bones, and the bones, being of the same substance as the descendants, passed the benefit on: health, sons, office and wealth, in roughly the order the texts give them. The Zangshu states the principle in a sentence; the later tradition built it into a system in which the fortunes of a lineage could be read from the graves of its ancestors and corrected by moving them, and a family in decline would exhume and rebury a grandfather in the hope of a better site. De Groot, watching the practice in Fujian in the 1880s, describes coffins kept in the house or in a temple for years while a site was sought and a date chosen, and describes the geomancer as the most consulted professional in the province, paid by the rich to find the lair and by the poor to confirm that the only ground they could afford would do.

The lineage and the landscape

Freedman, working from the records of Fujian and Guangdong, showed that the geomancy of graves was competitive. Good sites were scarce; the benefit of a site was thought to flow to the descendants of the person buried in it and to nobody else; and a lineage that secured a strong site was understood to have taken something from its neighbours, whose own graves now lay in its shadow. Families bought hills, fenced them, and sued over them. A grave placed so that it interrupted the vein feeding another was an act of aggression, and disputes over graves, their sites and the ridges that fed them ran for generations and filled the magistrates’ courts. The landscape of a southern county was, in this reading, a map of family claims drawn in tombs, and the geomancer was as much a surveyor and an advocate as a priest.

Prohibition, from the beginning

The state never liked it. Sima Guang, the great historian of the eleventh century, wrote an essay against the practice, complaining that families delayed burials for years while they hunted for a site and that a filial son should bury his parents promptly and let the ground be. Officials after him repeated the complaint under every dynasty: burials delayed, hills fought over, the poor fleeced by geomancers, the dead kept above ground for the sake of the living. The Ming and Qing codes fixed penalties for keeping a coffin unburied beyond a set time. None of it worked, because the belief was a belief about family, and the state had nothing to put in its place. The suppression after 1949 was the most thorough, and, as Bruun found, no more permanent than the others.

School · The compass at the grave

The luopan was a burial instrument before it was a house instrument. Its innermost rings give the twenty-four mountains for the orientation of a coffin, and the Compass schools developed their tables, the San He’s harmonies of dragon, water and facing and the Xuan Kong’s periods and stars, for the yin house first and applied them to the yang house afterwards. Needham’s history places the geomancer’s compass before the navigator’s: the south-pointing needle was reading graves for centuries before it read a course at sea. Skinner’s guide to the compass still describes several of its rings as burial rings, kept on the instrument by tradition and unused in a living room. A reader who meets the compass methods on this site is meeting the descendants of tomb-siting arithmetic.

The tomb as a valley in miniature

The tomb of southern China, Taiwan and the diaspora is shaped like the site it should have. A horseshoe of masonry curves round the burial from behind, the tomb’s own tortoise; two low wings reach forward from it, the tomb’s own dragon and tiger; between them, before the stone that names the dead, a small paved forecourt, the tomb’s own bright hall; and the whole faces down the slope to whatever water or open ground the site allows. Where the landscape could not supply the form, the mason supplied it. The Korean royal and gentry tombs Yoon describes do the same with an earth mound and a stone table, and the mound of a Chinese grave, however humble, is a hill built for the purpose.

Fig. 02 The tomb as the mason builds it, with the four parts the valley would have supplied. The wall curving round the back of the burial is the tomb's own tortoise; the two wings reaching forward from its ends, stepped in tiers, are its own dragon and tiger, the dragon on the tomb's left as it looks out and so on the right of the drawing; the paved court between them, with the offering slab in it, is its own bright hall; and the stone that names the dead stands across the axis where the mound meets the court. The tomb faces down the slope, to the water or the open ground the site allows. The shape is what is constant, not the size: graves run from a mound a couple of paces across to the terraced tombs a lineage builds across a whole hillside, and the drawing is set at a modest family tomb, about seven metres from wing to wing. De Groot drew the Amoy version of this plan in 1897 and called it the omega, for its shape; Freedman describes the same tomb in Fujian and Guangdong, and Yoon the Korean mound and stone table that answer to it. The drawing is a diagram of relations, not a survey of any one grave.

What the water does

The part of the burial reading that translates most directly to the living is the water. The texts want it to enter from the side, embrace the front of the site and leave slowly by a narrow mouth, the shui kou, between hills that seem to hold it back; the commentaries call those hills the locks, and count a mouth that is locked by two or three of them as the mark of a great site. A site whose water runs straight past and out is read as draining, and the reading is the same one a farmer makes: a valley whose water lingers is watered, sheltered and fertile, and a valley whose water runs out is a wind tunnel with a stream in it. The Shuilong Jing, the Water Dragon Classic, extends the reading to the plains, where there are no mountains and the watercourses do all the work; its drawings of rivers embracing and rivers cutting are the ancestors of the road drawings on this site.

Fig. 03 The water mouth, in the two readings the texts give it. A: two streams come down either side of the site, meet in front of it and leave through a narrow gap between two hills, the mouth; the water lingers, and the texts read the site as holding its qi. B, ghosted: one stream runs straight past the site and out across open ground; the water leaves as fast as it came, and the texts read the site as draining. The site itself, the lair and its arms, is the same in both. The mouth is the Form School’s test of a valley and one of the few of its rules that a modern eye reads the same way: a valley whose water leaves slowly is sheltered, watered and fertile, and a valley whose water runs straight out is a wind tunnel. The Zangshu’s sentence, that qi stops at water, is about this drawing.

What crossed over

When the tradition turned to houses, the yang zhai, it kept the drawing and changed the client. The mountain behind the grave became the wall behind the bed; the arms became the flanking walls, and then the arms of a chair; the bright hall became the clear floor inside the door, and then the open middle of a living room; the water in front became the road, the pond and the view; the water mouth became the door, the mouth of qi; the lair, where the veins rest, became the command position, the place a body is put where the room gathers rather than drains. The Huangdi Zhaijing made the transfer first, and the Ming compilations of the yang house made it systematic. Everything on this site about a room descends from it.

What did not cross over is the doctrine of resonance. A house cannot pass qi through anyone’s bones, and the yang house texts do not claim that it does. They claim that a house sited and arranged as a good grave would be is a house its inhabitants prosper in, which is a weaker claim and, in its practical parts, one the research can sometimes test: shelter behind, prospect in front, light, air, water in view. The rules about backing, view, light and water that the site labels tradition are grave rules in a second life, and the reader who wants to know why a bed should have a wall behind its head is looking at a mountain.

Reading the site

The yin house survives on this site as vocabulary and as drawings. The outside page reads a garden as a valley; the what is feng shui page draws the one shape at three scales; the history sets the burial classic in its time; the texts page says what the Zangshu does and does not contain; the sha catalogue is the list of what the sand and the water do wrong. The reader will not be asked to site a grave. The reader will be asked, on every room page, to put something solid behind a body and something open in front of it, and now knows where the request comes from.