Stage IThe beginning

One tree
standing

एक पेड़ खड़ा था — बाक़ी का कोई रिकॉर्ड नहीं

The nearest things,
and why none of them is it

25 September 2021. The earliest photograph of this land we hold. Kans grass over open ground, nothing worked, nothing planted. Seven weeks later the machines were on it.

Why this is still not the beginning

It is dated, it is this land, and it is before any of the work. It is also the earliest thing in two hundred and eighteen photographs, which is not the same as the earliest thing that happened.

Nobody has told us when the land was bought. So we cannot say whether this is what it looked like at purchase, or what four years of nobody touching it looked like. Until somebody who was there says which, it is a photograph of September 2021 and nothing more.

Scrubby, part-cleared ground at the edge of the farm. Undated. There is no tree in it, a transmission pylon stands near the centre, and a pickup with a legible plate is in frame — that plate would have to be obscured before this is published anywhere. It must never be captioned “before”. It is the nearest thing we have and it is not the thing. Nobody knows when it was taken, or whether it predates a single planting.

Four files that are not what they look like

They look like a historical aerial sequence of the farm, and they are not one. They are screen captures of someone else's satellite imagery, with a copyright line and a watermark burnt into the pixels, framing a whole village of which this farm is a fraction.

Unusable at any size, in any crop, under any caption. We do not treat them as if that sequence exists. A historical sequence would have to be sourced properly. Two possibilities are at the end of this page.

Stage IIThe work

A sequence,
not a chronology

From the air

Mostly bare ground, with rows already in it. The planting has started and the land has not caught up.

From the air

The two houses, still frames. Bare earth around them, bougainvillea already in on the boundary.

Neither of these is the beginning. Both are already the work. The earliest thing we hold is a farm partly made — which is its own answer about how much was never recorded.

1. 16 November 2021Dated by the photograph, not by a record

   The ground was read before it was changed

   The contours were kept rather than levelled. Levelling is easier with a machine, and it is what you do when the plan matters more than the place. Everything afterwards follows from that one decision. Where the water runs, where the beds sit, where the houses went.

   

   A tractor and a JCB, last light. This is the machine work that could have levelled the ground and did not.

   

   The cut, with somebody in it for scale. The bank follows the fall of the land rather than crossing it.

2. 16 November 2021Dated by the photograph, not by a record

   Water before planting

   A single borewell, and nine irrigation zones laid out to follow the ground rather than cut across it. Nothing could be planted at scale until this existed, which is why it must have come first, and why the weekly argument on this farm is still about which of the nine gets the water.

   

   Pipe on the ground before anything is in it. The same day as the earth was being moved.

3. Date not recordedNobody wrote it down

   Eleven species, not one

   The orchard went in over several seasons rather than all at once, following what was available and what survived its first summer. Eleven fruit species, not in blocks — a row may carry one for twenty metres and then change. Eight or nine of them flower between January and April.

4. Date not recordedNobody wrote it down

   The loop was closed

   Twenty-four vermicompost beds take what the farm grows and return it to the ground it came from. This could only come after there was something growing to return, which fixes it in the sequence even though nobody wrote down the year.

5. Date not recordedNobody wrote it down

   The systems arrived around the planting

   Food forest, multi-layer beds, turmeric under canopy, the dairy, the apiary inside the citrus strip, the mud houses on the high ground and a camping site below them. None of it arrived as a plan. It arrived one season at a time, which is also why the record has gaps.

Three dashed date slots down one edge, where there were five. Two of them could be sourced from a photograph. The other three still cannot, and that column is the real state of our own history.

One day
somebody photographed properly

1. 

   The rope

   Twisted grass rope. Every joint in the frame is tied with it; nothing is bolted.

2. 

   The frame

   Bamboo raised to a ridge and lashed. This is the whole structure — the rest is skin.

3. 

   The panels, flat

   Thatch woven into panels on the ground, where a person can work sitting down.

4. 

   Lifted

   A hired mobile crane. The one piece of the build that could not be done by hand.

5. 

   The apex

   Closed by hand, standing on it. Both houses were up by the end of the day.

We are not holding this up as proof of anything except that somebody was there with a camera. The farming is the part nobody photographed.

Stage IIIThe living landscape

What is here now,
and can be shown

The orchard in fruit. Kinnow is ready in January, which is also when most of the orchard is in flower.

The rise, and the track up to it. The planted beds run along the contour rather than across it.

Sowing by hand. No face is readable and nobody in it has been named or asked.

The Gir dairy. Its bedding goes to the compost beds; the beds go back to the orchard.

Stage IVThe ongoing experiment

The chapter
that is still open

अभी ख़त्म नहीं हुआ

Nothing has supplied an ending to this one. What follows is the current chapter, and it is the part worth reading.

August 2026. New frames going up, in the rain months.

Poles set, nothing on them yet. The multi-layer structure, at the stage where it is only posts.

A frame, standing open. The shade over it has not been made yet.

These are six weeks old. Naveen sent them from the farm in August. This stage is dashed on every other page because it has not finished — and this is what unfinished looks like.

1. The apiary collapsed

   Summer heat took it. Ten colonies are alive across about fifty boxes. What matters is the drawn comb standing in the empty ones. Comb costs a colony a season of work to build, and a colony moving into it starts a season ahead.

   Learned · the heat is the constraint, not the forage

2. The count went down

   The orchard was described for years by the number of trees that had gone into the ground. Then somebody walked the rows and counted what was standing, and the number came down. Planting tells you what went in. Standing tells you what is alive.

   Learned · prefer the smaller, truer number

3. Six systems have never been photographed

   Of thirteen working systems on these five acres, seven have a photograph and six do not. That includes the compost beds the whole loop depends on, and the apiary described above. Each one is named and left empty rather than filled with something else.

   Learned · a record kept late is still a record

4. The boundary has never been traced

   We do not have the real boundary in any form we own. Every position on the farm map is plausible rather than surveyed, and the map says so. Tracing it is a day's work that nobody has done.

   Learned · say plausible, never imply surveyed

5. The farm built something that can refuse it

   Because eight or nine of eleven species flower in the same window and the apiary sits inside the citrus strip, our software will not clear a spray date it cannot square with the apiary record. It does not warn and proceed. It refuses, and somebody has to clear it by hand or not at all.

   Trying · let the record constrain the behaviour

6. What has not been settled

   Harvest windows for most of the eleven species are in people's heads rather than written down. No date in this farm's history is recorded. Nothing here is finished, and we are still arguing about most of it.

   Open · and we will say so until it is not

Two things that
might exist