On the survey sheet the village is spelled one way. In the revenue records it is spelled a second way, which is the surveyor's spelling with one vowel altered by a clerk at some point before 1911. What people who live there call it is a third thing, close to the second, and not what either document says.
For a century that did not matter very much, because a human being read the record and knew which place was meant.
It matters now, because the records are a database and the database matches on strings.
I have been tracing this particular village through documents for a different project and the trail is instructive in a way I did not expect. The surveyor who produced the sheet was working in the 1870s, transcribing a name he was hearing, into a script and an orthography that had no settled convention for the sound in question. He made a defensible choice. It went onto the sheet, and the sheet became the authority, because a map is a thing you can hold and a spoken name is not.
Every subsequent document copied the sheet. The revenue record copied it with an error. The census took the revenue record. The electoral roll took the census. The post office took something. And when the land records were digitised, the digitisation took the revenue record, because that is what the revenue department had.
So there is now a chain seven documents long descending from one man's ear in the 1870s, and the village's actual name is not in it anywhere.
The tidy version of this observation is that colonial knowledge production imposed its categories, and I have written before about why I distrust that move — not because it is false but because it is available in advance of any evidence and explains every case identically. So let me say what this case actually supports, which is narrower and stranger.
Nobody imposed anything. The surveyor was not asserting authority over the name; he was writing down what he heard so he could finish the sheet. The clerk who altered the vowel made a slip. What gave the spelling its power was not intent, it was durability: the written form persisted and the spoken form did not need to, because everybody local knew the place and nobody local was consulting the sheet.
Then the sheet stopped being a description and became the operative record, and it did so in stages, each of which was somebody making a sensible administrative decision.
There is a live consequence and it is not sentimental. Three households in that village have inheritance documents that spell it the third way, because the deeds were written by a local writer who wrote what people say. In the digitised system those documents do not match the parcel records. Somebody must now certify that the two names denote one place, which is an affidavit, which is a fee and a trip to the district headquarters, and which — I have checked — has been required of about forty households across the four villages I looked at.
A century of orthographic drift, resolved daily at the expense of the people who never spelled it wrong.
I do not have a proposal here and I distrust the instinct to produce one, because the version I keep drafting is a mapping table of alternate spellings, and a mapping table is exactly the sort of thing that gets built once, incompletely, and then becomes the new authority. It would fix the forty households and create a different problem for somebody in 2090.
What I have instead is a note in my own files: whenever I cite that village, I write all three names. It costs a line. It is not a remedy for anything. It is the only thing I can do that does not require me to decide which spelling is correct, and after four years of this I am no longer sure that question has an answer that a historian is entitled to give.
