Monday, November 22, 2010

Monday Mailbox: Indexing Place Names

Indexing place name abbreviations“When census enumerators wrote ‘Ind.’ and the person is named Charlotte Smith,” wrote Brad, “and yet the birthplace is transcribed as India, I just have to laugh.”

Brad, you may be interested to learn that Ancestry.com and FamilySearch handle abbreviated place names differently.

Ancestry.com indexers enter abbreviations exactly as written. It is Ancestry.com themselves that erroneously expand “Ind.” to “India.” I understand they have fixed these. If you can find any surviving errors of this type, I would be interested to learn of them. Because indexers were asked to enter the information exactly, Ancestry.com can go back and efficiently reprocess millions of records. Preserving the exact contents of the fields is a best practice in the archival world.

FamilySearch Indexing, on the other hand, asks indexers to interpret place names, correct spelling, and expand abbreviations. This generally produces better results when indexers have sufficient contextual knowledge. But misinterpretations occur, such as picking a far away place in the same state instead of a really close place just across a state line. Because exact contents of records are not captured, errors can not be corrected without re-indexing.

-- The Insider

8 comments:

  1. All transcriptions are scary.....Find the actual document! It is my understanding that the census we see is more likely the THIRD copy that has been made. No wonder the census is not accurate, given the other issues just with taking it originally.

    ReplyDelete
  2. Here's another one for you: when people use the abbreviation NL in place names in their genealogy program to denote a place in the Netherlands, Ancestry expands this to Newfoundland and Labrador. See this example.

    ReplyDelete
  3. Seeing how I am a 6th generation Hoosier I have run across more than my share of "Ind" = "India" but my latest pet peeve with Ancestry.com's locations is when you type in a location and it fills in the rest... handy .. but only if correct! It has Indianapolis listed in Hamilton County. This is technically correct... for literally a few small blocks and just recently due to the sprawl of the city. However the entire county of Marion is also city of Indianapolis.


    For all those that are not familiar with the area and automatically assume that Ancestry knows what it is doing has Indianapolis, Hamilton, Indiana listed as their location. Think of the hundreds of hours of useless searching for sources only to discover that Indianapolis is in Marion County!

    [This doesn't even address the issue that there is a city located south of Fort Wayne in Indiana that is call Marion. Ancestry also switched Marion County, with Marion the city located in Grant County.]
    I have sent a message to Ancestry but... no reply.

    ReplyDelete
  4. Dear Yvette,

    I clicked on your link and couldn't find the example. Can you post a link to the image?

    -- The Insider

    ReplyDelete
  5. AI you said, "Ancestry.com indexers enter abbreviations exactly as written. It is Ancestry.com themselves that erroneously expand “Ind.” to “India.” I understand they have fixed these. If you can find any surviving errors of this type, I would be interested to learn of them."

    Here is what I found concerning those tabulated and/or indexed as born in India, in Ancestry.com's US Federal Census indexes.

    1860--9,172. A search for those who resided in Indiana turned up some very strange indexed entries for birthplace, such as "Germany;India" which was mistranscribed from the enumerator's writing "Hesse-Darmstadt, Germany," with Germany on the next line but clearly enclosing the long place-name in brackets, and the birthplace of the next person written as "Ind." The "W├╝rttemberg;India" entry is quite interesting. Seven people in De Kalb Co., Indiana were indexed as born in "Indi," which was returned as part of my exact search for people living in IN, b. in India. However, these persons were not returned in an exact search for persons born in "Indi," although others residing elsewhere in IN as well as in TX were returned in the results. A few tabulated as born in Bombay and Calcutta were in search results for those b. in India.

    1870--28,152, of whom 9,604 resided in Indiana

    Another famous Ancestry.com place-name anomaly is tabulation in the 1850 US Census of persons b. in "Ia" as born in "Iowa." There are 190,039 entries for persons indexed as b. in "Iowa," begining in 1751. Those born in "Iowa" 1751-1771 all lived in Indiana in 1850. Of all those born in "Iowa," 116,290 (61.2%) lived in Indiana, and 48,241 (25.4%) lived in Iowa.

    A search for those enumerated for 1850, indexed as born in "India," turns up some very interesting results. One Mary Divis, of Manchester Twp., Morgan Co., OH, who was enumerated as b. in "O." is tabulated as b. "OI" and returned in search results as b. in "India." The many others b. in "O." on the same page were tabulated as b. in Ohio. Those tabulated as b. "India;Tennessee" were b. in Indian Territory. One person indexed as born in "Massachusetts" is inexplicably returned in the "India" birth-place results, as is one tabulated as b. in "Indiana." Two persons tabulated as b. "Inde," enumerated as b. in "Ind." were also returned in the "India" birth-place results. One person tabulated as b. in "Pi," enumerated as b. "Pa.," was also strangely returned in the "India" birth-place results. In an exact search for "Pi" birthplace, she is returned as the sole result, but those tabulated as b. in Calcutta or Bombay were not returned in an exact search for those places (though listed in the "India" results).

    For the 1860 and 1870 results I could only do spot checks, but all I looked at that were straightforward were enumerated as born in "Ind."

    As Genealem said, "All transcriptions are scary . . . ." and vagaries of indexing conversions on Ancestry.com are often weird.

    Let no one mistake a transcript for a "record."

    ReplyDelete
  6. I'm sorry, it wasn't an example of a transcription error but of another type of auto-correct error that Ancestry made. There is no image, but if you check the birth place of the record in the link, you'll notice that Winterswijk (a town in the eastern part of the Netherlands) is suddenly located in Newfoundland and Labrador, Canada.

    ReplyDelete
  7. One cannot begin to imagine how badly the records for Cornwall (UK) are transcribed and indexed, when done by those who are not Cornish. People assume that because Cornwall is now part of England, that indexing would be easy for an English speaker. But most of the place names and many of the family names are Cornish (a Celtic language similar to Welsh)in origin and the transcriptions often bear little resemblance to the actual name.

    An English census enumerator or English vicar wrote what they thought that they heard (the pronunciation is unique to the Cornish people) and only with a knowledge of Cornwall with its language and dialect can one try to think of what an Englishman might have written in order to perform a search.

    So it reinforces the point about how vital it is to have local knowledge to accomplish reliable indexing or transcriptions.

    ReplyDelete
  8. I've found relatives who were born in India (to a missionary father and his wife working there) listed as being born in Indiana on ancestry in passenger lists for San Francisco. The only thing I could potentially correct was name which was already correct and not the place of birth which was wrong, so as far as I know it's still wrong. It doesn't matter what someone's name is - someone else wrote that a Charlotte Smith was born in Indiana was listed as being born in India and he thought that was funny. My Charlotte (Svensdotter) Smith was born in Sweden and first married a Smith in Minnesota. After he died she married my Swedish great grandfather and my grandmother was born. The last name of my relatives born in India is Larson. They're Americans of Swedish descent born in India. The moral of the story is you can't always tell if a birthplace is correct based on surname.

    ReplyDelete