Transcription versus original: the error that cost me a year
August 2024 · 3 min read
In my second year I built an entire branch of the tree on a birth year of 1848, taken from an index entry. The index was a transcription of a civil register, typed by a volunteer decades ago, and it said 1848 very clearly. For about a year everything I found seemed to fit that date, with the usual small inconsistencies that I explained away. Then I finally paid to view the original register page, and in the clerk’s hand the year was unmistakably 1843. His 3 had a flat top and a long tail that looked like an 8 to someone working fast. Five years was enough to make the man I had been researching a different man, with different parents. The branch came down.
The lesson is the obvious one and I am still annoyed I had to learn it in person: an index is a finding aid, not a source. It tells you where to look. It does not tell you what the record says. Every index is at least two steps from the event: someone wrote the original down, and someone else read that writing and typed what they thought they saw.
What I do now with any index hit:
- View the original image before writing anything in the tree, and cite the image rather than the index.
- Compare any doubtful digit or letter with the same clerk’s other digits and letters on the same page. A hand is consistent within a page; the 3s will all have that flat top.
- Read the entries above and below. Registers were kept in order, so the dates of the neighboring entries bracket the one in question.
A second case taught me the same thing from a different angle. A surname I was looking for did not appear in a parish index at all, for twenty years in which the family was definitely there. When I browsed the register page by page, the entries were present under a different initial letter, because the clerk’s capital had a flourish that the transcriber read as another letter entirely. The index was not wrong, exactly; it was faithful to a misreading. Browsing took four evenings and found eleven baptisms.
I keep the wrong 1848 branch in a separate file, labeled, as a reminder. It is forty pages of careful work on a stranger, and it is the most instructive thing in the whole archive under the stairs.