Randy Seaver has more AI in store for us for this week's Saturday Night Genealogy Fun.
Your mission, should you decide to accept it (cue the Mission: Impossible! music) is:
1. FamilySearch Full-Text Search continues to add databases and searchable images to their collections. This is a gold mine, especially of land, probate, and court records.
2. Pick one or two of your ancestors and see what you can find on FamilySearch Full-Text Search about them.
3. Share your Full-Text Search find(s) in your own blog post or in a Facebook, SubStack, BlueSky, or other social media post. Leave a link to your post on this blog post to help us find your post.
The results for most of my searches for my ancestors (and remember, an ancestor is someone from whom you descend, not a collateral relative; since there is no such thing as an "indirect ancestor", the term "direct ancestor" is redundant at best and nonsense otherwise) either had far too many results to look through, when I searched for just a surname, or no results at all, when I searched for full names in quotation marks to control the number of irrelevant results. My names are not extremely common ones, such as Smith and Jones, but they are common enough that a blind search with no index produces far too many results to slog through. My grandmother's name was Anna Gauntt, with no middle name that I have ever discovered, so I searched for "anna gauntt" and learned that the AI provides results with something between your search terms. I was able to rule out all of those Anna Gauntts, because they either had middle names or initials or were not in the correct locations for my grandmother. I abandoned my ancestors and searched for some of the unique surnames that I am doing research on.
My search for Gudapel, a name which has been used by only four people in the history of the world, produced two results. Both had headers that read "History Records 1800–1902, Diaries 1800–1902 | New Hampshire. Genealogies 1978–1982, Society Records 1978–1982 | Maine. Genealogies 1978–1982, Society Records 1978–1982" (truncated on screen, but visible when I moused over the link). If I were searching for a name that I did not know as well, I might have ignored these results, because I know the family was never in New Hampshire and Maine and would have no reason to appear in genealogies for those area. Because I do know the name, I clicked on the links, hoping that maybe there was something from the 19th century, and discovered that the title was not particularly accurate. Both links went to the same book, a 1941 Houston, Texas city directory, which did include the name of Geo. [George] Gudapel on two different pages. While city directories can be classified as history records, the directory in question does not fall in the years of 1800–1902. I'm not sure if I already had these directory listings.
I next searched for McStroul, a name which my aunt's grandfather created when he naturalized as an American citizen. His original name was Moska Leib Strul. He asked to have it changed to Leo Martin McStroul when he became a citizen. The name McStroul belongs only to his family; when I find it, I know it has to be the right people. When I entered it in the keyword field, I had 42 results.
I searched for McStroul when Randy did a previous FamilySearch Full-Text Search challenge. At that time I had two results, both in my aunt's great-grandmother's will and probate. The name appeared because my aunt's mother, who was the granddaughter of the deceased, was listed under her married name. I noted at the time that the name actually appeared three times in the document, but only two were identified by the search. This time the search picked up all three occurrences.
All 42 results for McStroul that Full-Text Search found were correctly read. A couple of the given names were misread by the AI (such as Geo instead of what actually said Leo). Many of the documents were ones that I have not previously found, such as articles in a Kingman, Arizona newspaper and naturalizations in Milwaukee, Wisconsin where the McStroul family members appeared as witnesses. So these are all useful in researching the family and are nice discoveries.
I tried searching for Szocherman, a name in my family that I believe more and more may be unique to that branch of my cousins. I had no results. But when I searched for Socherman, a spelling which some family members have adopted, I found many results, almost all of which were not for my cousins. Amusingly, one that was my family was misread by AI and actually does say Szocherman.
After all of that, I tried to find a collateral relative in the database and searched for "frederick dunstan" in quotation marks. Again the AI provided results with something between my search terms, and there were far more Frederick Dunstans than I had anticipated, more than was practical to look at. I restricted the search to New Jersey and ended up with only four results, three of which were for Frederick C. Dunstan in Burlington County, which is exactly where my great-grandmother's brother should be found. I believe this is the right person, so it appears he had a middle name, which I previously did not know, that started with C. All three results are from deed indices from the early 1920's. That does provide me with information about him.
Before I restricted the search to New Jersey, some of the results were for a Frederick Dunstan in Allegheny County, Pennsylvania. I looked at them bcause that isn't horribly far from New Jersey and found that they were for a Frederick Dunstan from Combe Martin, England, which is in Devon County. My Dunstans were from Lancashire, so I knew this was not my guy, but ironically, his wife's name was Jane. Jane Dunstan is my great-grandmother and the sister of Frederick Dunstan. I did find that entertaining.
