Dividing Lehman and Lehmann
I want to devise method(s) that distinguish Lehman and Lehmann from one another. I was thinking in terms of AI and one method being determining different naming conventions for Lehman and Lehmann. Does anyone have ideas around this?
Answers
-
I fear that this is, in a practical sense, impossible. Names get recorded by people who prefer one spelling or the other (and may not speak the same language as the person they are recording), families change how they spell their names generation to generation (sometimes drastically), and sometimes people just change their mind and spell their name differently than how it was given to them.
While a computer learning algorithm could tell you what would best match the patterns it was trained on, it's more likely to give you false confidence (and false positives) than actual insight.6 -
And the further back you go the more fraught with complication any attempt at division is going to be. Prior to the early 1900s, there was no such thing as standard spelling for names or anything else. If anything, if you are working out a couple of different families you will likely need to combine spellings, not separate them. You'll need to look for every possible spelling then use other means to separate them into families.
But there are probably many examples of modern spelling changes that you have to account for to not mistakenly assume a lack of relationship. For example, one of my father's brothers decided he was going to spell his last name Collette. That is how all his children and grandchildren spell it. But he is still my uncle.
5 -
To add to Gordon's comments, one of my ancestors, from Brittany (France), had the surname "Michel." When the family settled in the southern USA, the name began to change, partly due to Civil War paymasters. My 2nd GGF eventually used "Michael," while some of his siblings used "Mikell." They are all still my family.
5
