Swedish patronymic names and mistakes in extractions and copying of sources.
LegacyUser
✭✭✭✭
Gil Mariana Herrey Torriente said: From a 100% Swede who speaks Swedish, and does daily research, here is something to watch out for and something that needs change:
Whoever does the extractions and matching of sources to people for Sweden (and probably Norway and Denmark, too) makes big mistakes with last names. Up until the 1890's the patronymic system was used, but it is completely ignored in so many cases. Anders Bengtsson's children born in the 1700's are Andersdotter (Anders' daughter) and Andersson (Anders' son), NOT Bengtsson. Peter or Per or Petter Larsson's children born in the 1800's are Petersson (or Persson or Pettersson) or Petersdotter (or Persdotter or Pettersdotter)all variations of the same name, but NOT Larsson. So for those who are doing research and come across all the mistakes in the source section or even in the old extraction of names, and don't know the system, they will copy down the wrong information. I am grateful for the addition of sources, but the persons who copy information from household examination records, where there usually is no last name written down for the children, cannot assume that their last name is the same as the father's. Having said that, when Sweden changed all that in the late 1800's, some children chose to keep the patronymic name, some chose to have the father's last name, in the same family, so that is very confusing.
Going back to old extraction of names, and submission of names a long time ago, such in the 1960's when everything was submitted on family group sheets, whoever was in charge at the church offices, did not allow Andersdotter, Larsdotter etc, They made the submitters change all girls' names to Andersson, Larsson etc, so when researchers come across that now, they think that is correct, which it is not. ( This is also the same for Norway and Denmark) I change those all the time to what it should be. Again, a daughter, dotter, could never be a son. She was always a daughter, dotter. Datter, in Norway and Denmark.
Whoever does the extractions and matching of sources to people for Sweden (and probably Norway and Denmark, too) makes big mistakes with last names. Up until the 1890's the patronymic system was used, but it is completely ignored in so many cases. Anders Bengtsson's children born in the 1700's are Andersdotter (Anders' daughter) and Andersson (Anders' son), NOT Bengtsson. Peter or Per or Petter Larsson's children born in the 1800's are Petersson (or Persson or Pettersson) or Petersdotter (or Persdotter or Pettersdotter)all variations of the same name, but NOT Larsson. So for those who are doing research and come across all the mistakes in the source section or even in the old extraction of names, and don't know the system, they will copy down the wrong information. I am grateful for the addition of sources, but the persons who copy information from household examination records, where there usually is no last name written down for the children, cannot assume that their last name is the same as the father's. Having said that, when Sweden changed all that in the late 1800's, some children chose to keep the patronymic name, some chose to have the father's last name, in the same family, so that is very confusing.
Going back to old extraction of names, and submission of names a long time ago, such in the 1960's when everything was submitted on family group sheets, whoever was in charge at the church offices, did not allow Andersdotter, Larsdotter etc, They made the submitters change all girls' names to Andersson, Larsson etc, so when researchers come across that now, they think that is correct, which it is not. ( This is also the same for Norway and Denmark) I change those all the time to what it should be. Again, a daughter, dotter, could never be a son. She was always a daughter, dotter. Datter, in Norway and Denmark.
Tagged:
0
Comments
-
Heidi Kuosmanen said: This same problem appears in Finnish record extraction, because our records were in Swedish and we also had that patronymic system widely used.0
-
m said: On my Swedish LDS lines I do as little as possible due to this reason and high level of difficulty.0
-
Gordon Collett said: But the Swedish records are so great! And all original microfilms are available for free, some on FamilySearch and all at the Swedish archives:
https://riksarkivet.se
Once you figure out the few quirks in the older FamilySearch databases there are a lot of records that are very easy to use. And once you get familiar with the clerical household surveys, some of which are being indexed by FamilySearch and MyHeritage, you can follow your family practically year to year through history. It is really fun and the number of sources that can end up on a person can be huge.
I would encourage you to find and sit down with someone very familiar with the Swedish sources and get more familiar with these.0 -
Robert Wren said: And the Swedish Parish records are also available at FHC's (free) through the Arkivdigital program. It is very complete and generally quite 'easy' to use, with the resident Parish the key.0
-
m said: Thanks, for the encouragement!
Actually, due to so many DNA matches on Swedish lines, I am helped greatly by the fact that the DNA matches' trees and their documentation happen to line up so very exactly with my tree and documentation. (Even the surnames. So lucky!)0 -
Lundgren said: I can also confirm the issues you are experiencing and share your frustrations, as I have done a fair amount of research on my Swedish lines.
While I don't have a solution for you, I do have a few suggestions that I have found helpful.
When searching for females with patronymic, use a wildcard in the surname like this:
https://www.familysearch.org/search/r...
If you are searching for a birth record, it is possible that the birth record was correctly indexed w/o a surname. To find them do NOT include any surname in the search.
You can add the parents names like this:
https://www.familysearch.org/search/r...
The first few results on that page are records that you will most likely not be able to find if you add a surname. You *might* find them down lower in the results if you use a surname with a wildcard.
To narrow the results you CAN add the parents surnames, as they were probably on the record. I did not here as I was not after any record in particular.
Making it easier to find records w/o surnames is on our back log of things to do.0 -
Gil Mariana Herrey Torriente said: Great comments. I really wasn't asking for a solution, but pointing out a problem, especially for the smart computer engineers and experts who work for Familysearch, so that they can be aware of the problem and maybe fix it. And also to help others who are doing research but are not aware of this issue. I grew up in Sweden, speak the language fluently, and have been doing research since I was a teenager. I am now 68 and have at least 30,000 names that I have found and researched. Now, I fix things that are wrong, spellings, patronymic names, places, etc. I match duplicates and also separate names that are NOT duplicates, and much more.
So many times, I found that names have been entered many times and are still being entered. Often, a person is already in Familysearch, but with a patronymic name, when there is actually a real last name, and it doesn't come up as a match, The indexer or extractor used the patronymic name, when there actually is a real last name in the record, so when i make corrections, I mostly use both, such as Andersdotter Lindgren, or Hansson Bergman so that nobody enters the name again. And I use both as last names.
For many years, we couldn't find my paternal line, with the last name Herre. They lived in one parish, the records had burned and we couldn't figure out where they came from before. It turned out that all the birth records in the parish that they came from had been extracted many years ago, and my ancestors had lived there for generations, but NOT ONCE had the extractors copied the real last name, which was there in every birth record. So for example, the children of Johan Herre, were all ....Jansson; (short for Johansson) in the records instead of with the last name Herre. Herre was never copied as a last name, even though it was in the records every time, No wonder it was hard to find, but once I figured that out, it opened up thousands of names of my ancestors and their families.0 -
Glenn P said: Thank you all for your original post, and the added comments, insights, suggestions, experiences, and added work to make the Tree accurate. It is very very much appreciated, for those getting their heads around Swedish/Nordic lines.0
-
Gordon Collett said: As far as the search engine goes, it completely ignores whether a surname ends in -sen, or some variant, or -datter or some variant. If you are searching for Anna Larsdatter, you don't need a wildcard because the search engine will find all the Anna Larssons and all the Anna Larsdatters all by itself which I think is a great piece of programming.
Wild cards don't help for the real problem. Searching for Anna Lars* will not help when the index incorrectly has Anna Larsdatter the daughter of Lars Anderson listed as Anna Anderson.
I do a huge number of searches in the Norwegian historical database by putting in just the child's first name and just the father's first name, along with places and dates to narrow the search. It works just fine. I'll frequently get the red error banner on the the top that says "try entering a last name" but I just ignore that, hit the search button again, and then will get the results.0 -
Gordon Collett said: Unfortunately, posts here on getsatisfaction are rather transitory. After they roll off the first page they may never be found again. Also, these posts tend to come to the attention of just the most active and experienced users of FamilySearch and Family Tree. We're also the ones who love to share advice!
With your extensive background, what you might want to consider is evaluating the Research Wiki pages for Swedish records and improving them. Like any Wiki, anyone can edit these pages. You do have to click a link to request editing privileges but when I did that, it was granted within a couple of days.
Here is the one for the indexed birth record:
https://www.familysearch.org/wiki/en/...
Currently, the sections "How Do I Search This Collection?" and "I Can’t Find the Person I’m Looking For, What Now?" are pretty skimpy. And nowhere does it mention the problems in the database which you discussed and which I would assume at this point cannot be corrected, just dealt with.
By editing the page, you will contribute to a permanent reference which, I hope, new researchers are taking advantage of. The Research Wiki has been given a more prominent place in the help menus, is there the Learn More links in the database descriptions lead, and is very prominently displayed on the country specific search pages:
This would be a way for you to spare new researchers the troubles you have had and help them learn the best way to use the databases to get the best information in Family Tree.0 -
Gil Mariana Herrey Torriente said: Thanks, Gordon Collett, I was told by Familysearch, when I complained about the problem, to write on this page and that the experts check on this page frequently. But I thank you for all your information. Much better than what I got from the Familysearch. I will look into Wiki. And Glenn, I am happy to help anybody anytime.0
-
Lundgren said: Gordon,
Thank you for your comments, testing and compliments on our system!
The solution to make name searches work is a bit different what you have described.
It does not target patronymic names with son/sen dotter/datter directly.
It is a more generic solution intended to target more than just the patronymic cultures. As such it will appear to work as you have described in many cases. In others, it does not work as well.
Consider these given name Abjørn in searches in Norwegian records:
https://www.familysearch.org/search/r...
Now use Abjørnsen as the surname:
https://www.familysearch.org/search/r...
And now Abjørndatter as the surname:
https://www.familysearch.org/search/r...
And now the wildcard instead:
https://www.familysearch.org/search/r...
You will see that the results from all three surname searches are very different.
The method that we use to deal with sen/datter (son/dotter) is not perfect, and cannot currently cover all of the cases. The wildcard in place of the ending is useful tool that can help.0 -
Gordon Collett said: Thanks for the information.
Looking through your sample searches, however, there are a lot of strange results I wouldn't expect to see there, even with FamilySearch's fuzzy search on names. I think part of the trouble is that Abjørn is either a very uncommon name or just downright a misspelling for Asbjørn or Anbjørn. So I'm going to set up a different set of searches. Now, since Anbjørn is a girl's name, it would never be found in a patronymic. I'm also going to set up the search the way I usually do, which is through the specific collection Norway Baptisms, and limit the geographic area to a single county. This gives these results:
With Asbjørnson:
https://www.familysearch.org/search/r...
With Asbjørnsdatter:
https://www.familysearch.org/search/r...
Scanning fairly quickly through these, it looks like the results are the same. The fuzzy search pulls in a lot of Bjørnsons and Bjørnsdatters.
Now using the asterisk.
With Asbjørns*:
https://www.familysearch.org/search/r...
There is a big difference in the results which looks to be due to the apparent fact that using the wildcard forces an exact search on the portion of the name that is there. All the Bjørnsons and other fuzzy variants are gone and the results have dropped from 1,458 to 662.
Just for fun, let's use Asbjørnsd*:
https://www.familysearch.org/search/r...
Again, the exact search function comes into play, giving only 3 results. This means that there must be 659 Asbjørnsons. Statistically, one would expect a Asbjørnson/Asbjørnsdatter split of roughly 50/50 or about 331 of each. This is, of course, one of the points of the original post, that a lot of patronymic names were entered incorrectly in the database.
Also, this means that using the wildcard, one will miss all the typical spelling variants. For example, Asbiørnsd*:
https://www.familysearch.org/search/r...
gives 15 more results.0 -
Adrian Bruce said: "using the wildcard forces an exact search on the portion of the name that is there"
Gordon, that's an important point for all of us. I'm sure I knew that. At least, I think I did! But I'm also certain that if asked to explain things, that wouldn't have been on my list.0 -
Gordon Collett said: I've found something very interesting in the database of records for Stord, Norway, this morning in regards to this incorrect patronymic problem. I was looking up something there today and all the patronymics are gone:
This is how the records were originally indexed and solves the problem which was originally reported here.
This is a great first step, but I hope it is only a first step because it introduces a new problem that is going to really confuse people searching these records until they catch on. If you include a last name in the search, one will never get any results:
If I look for the top name in the above image using Knud Andersen with this search:
https://www.familysearch.org/search/r...
I get no results:
I have to remove Andersen as a last name and include Anders as his father's first name:
https://www.familysearch.org/search/r...
to find him:
or do a first name only search:
https://www.familysearch.org/search/r...
The trouble with the first name only search is that when the site it busy it will trigger the timeout warning "Please try again later, including a last name may help" (or however that is actually phrased.)
It should be encouraging for all who have posted here about this problem and under other discussions here about it that steps are being taken to address these patronymics. But, again, I do hope this is just a first step towards getting the correct patronymic applied.0 -
Lundgren said: You found that very quickly!
The removal of the assumed surnames on the older Scandinavian records started last night and is still going through the system. (This does not make any changes to the FamilyTree, just the records.)
The update should finish today barring any problems.
You are correct that that you will generally not find a record that does not have a surname if you supply one in your search. This is an issue across all of the records without surnames, and we have that on our backlog to address.
You should be seeing fewer "Try again" prompts now than you were seeing in the past, but they still do happen.
Thank you!0 -
Juli said: If indexers correctly follow the instructions for the Hungarian civil registrations, then birth records have just the child's given name, no surname (for example https://www.familysearch.org/ark:/619...). If a searcher then follows what little instructions there are for Search - Records (i.e. the frequent "try that again with a surname" message), then these records will never show up. Damned if you do, damned if you don't.0
-
Gordon Collett said: The Norwegian Archives Digitalarkivet transcriptions of records are very strict in that nothing is put in the transcription that is not in the record other than calculated years at times so last names of children are rarely found in the birth records. I'm very familiar with not using last names to search from that system.
Like all search systems, it is very important to first get familiar with the data set to be able to search efficiently. Unfortunately it seems a lot of newcomers to FamilySearch skip that step. I hope this change does not cause problems in people giving up too soon in finding records.
As the teams work further on these databases, will part of the work be to fix the indexes in which everything was done in capital letters? Or will we just have to live with that?
Here's an example:
https://www.familysearch.org/search/r...0
This discussion has been closed.