Reading Feng Shui in Water Country

Translator's note. In a water town — Shaoxing in Zhejiang, or Venice — the houses are surrounded by canals and people travel by boat. Read by the ordinary sand-water rules, such a place is a disaster: water on every side, and water in opposing palaces. The people who live there are prosperous. This short piece gives the correction, and it turns on one idea: a grid of water everywhere is not water, it is the baseline.

A water town inverts almost every assumption a feng shui reader carries in from dry land. The house is not on a site with a stream running past it. The site is a strip of land inside a network of waterways, and the waterways are the roads.

Apply the ordinary rules literally and the verdict is catastrophic. Water lies to the east, the south-east, the south, the south-west, the west, the north-west, the north and the north-east. Every palace holds the same element. Worse, every palace has its opposite palace holding the same element as well. This is the condition called opposing palaces holding the same water (对宫同水), and in a conventional reading it is about as bad as a site can get.

And yet the people of these towns are prosperous, healthy, and long-established. Some of the most successful families in Chinese history come from exactly this landscape.

The contradiction is not a defect in the rules. It is a defect in how the rules were applied — specifically, in what the reader chose to call water.

Why the standard reading fails here

The error is one of scale, and it is the same error that makes a reader condemn a whole city block for having no empty land.

On dry land, water is an interruption in the ground: a river, a pond, a ditch, a road. Because it is an interruption, it is a local feature, and its position matters — water in the east means something because the east is otherwise land.

In a water town, water is not an interruption. It is the substrate. Every plot is bounded by it, every route runs along it, and its level is the same everywhere because the whole network is connected. A feature that is identical in all eight directions cannot discriminate between them.

A variable with the same value everywhere carries no information. That is the whole correction, and it explains why the rules had to be adapted rather than abandoned.

The water grid is the zero plane

Treat the canal network as the zero plane — the flat baseline against which sand and water are measured, exactly as flat open ground is the baseline on dry land.

This is not a figure of speech. The reasoning is the same in both cases. On dry land, a field of wheat or a market garden is not read as a feature; it is the ordinary surface of the earth, and it does not tilt the reading in any direction. In a water town, the water network plays that role. It is the field. Treat it as the field.

Two consequences follow.

First, the canals do not count as water for the purpose of the eight-direction reading. They are not in any palace; they are under all of them. A reader who counts them will produce the absurd verdict described above, and will then have to invent a theory to explain why the actual town contradicts him.

Second, ponds that are continuous with the network are treated the same way. A pond fed by the canals, at the same level, part of the same body of water, is also zero. It is not an independent feature and cannot be assigned a direction.

What remains, once the grid is set to zero, is the small set of things that genuinely interrupt it.

Crossings are the water

The features that do count are the crossings — the places where waterways intersect, where four branches meet at a junction, where a junction opens into a wider pool.

A crossing is a local concentration of water at a specific point, and that point sits in a specific palace. Read it the way you would read a pond on dry land: identify the palace the crossing occupies, and judge it by the ordinary rule for water in that palace.

Two of these deserve specific mention because the tradition treats them as fatal. A crossing sitting on the chen position, and a crossing sitting on the hai position, are both read as indicating a short life. These are the two positions the school treats as sensitive for longevity, and they apply here exactly as they apply elsewhere.

The opposing-palace condition has not disappeared, it has merely moved down a level. If there is a crossing in the south-east and another crossing in the north-west, the two still form an opposing pair, and the pair still sits across the origin point. The reading is the same as on dry land: this combination is severe. The difference is that the pair is now made of two specific junctions rather than of the general presence of water, which means the reader has to find them rather than assume them.

This is the practical gain. In a town where everything is water, the only things that can be located are the junctions. So the junctions are what you read.

Inside the house

Inside the building, nothing changes. The interior is read in the ordinary way, and the ordinary assignments hold.

Kitchens are sand. Electrical equipment — refrigerators, televisions, computers, the meter cupboard — is sand. Bathrooms are water. A fish tank is water. The house inside a water town is not a special case, because indoors the water network is not present at all; what is present is the same set of objects that would be present in a house anywhere else.

The one thing to keep in mind is that the interior reading and the exterior reading use different baselines. Outside, the canals are zero and only the crossings count. Inside, the ordinary object assignments apply. Confusing the two produces either a house with no features at all or a canal network read as a moat.

The exception: graves not surrounded by water

There is one case where the grid does not get set to zero, and it is important enough to state separately.

For a grave in a water town, the rule holds only if the grave is actually enclosed by the water network — water on all sides, part of the same connected system. In that case, read it exactly as a house is read, with the grid as zero.

But if the grave is not enclosed — if it sits at the edge of the network, or on a tongue of land with water on two or three sides and dry land on the others — then the network is no longer a baseline. It is a feature, and it must be read as water in whichever palaces it actually occupies.

The reason is the same reasoning as before. Zero is not a thing, it is a relationship: a surface that surrounds the site equally in all directions. A grave at the edge is not surrounded equally. The water there has a direction, and a direction is information.

Getting this wrong is the most common error in water-country readings. A grave on the edge of the canals, read as though it were in the middle of them, will be judged inert when it is in fact sitting on a strong water configuration — and the resulting prediction will be wrong in whichever direction the water actually lies.

The general lesson

Water country is worth studying even for readers who will never work in one, because it isolates a principle that is easy to miss on ordinary ground.

Sand and water are not substances. They are roles, assigned by comparison to a baseline. Change what counts as the baseline and the same physical landscape yields a different reading — not because the rules changed, but because the reference did.

A river beside a house is water. A river that surrounds the house on every side, at the same level, as part of a connected network, has stopped being water and become the ground. The physical fact is identical. The role is not.