A family tree from public records

This week I built my family tree. I started with very little: my grandparents’ first names, one maiden name I had only ever heard spoken, and the family memory that a grandfather had fought in the Warsaw Uprising. Four days later the tree holds just under a thousand people, counting brothers, sisters and in-laws, and on one line it reaches ten generations back, to a farming couple in a village east of Warsaw in the 1760s. An AI assistant did most of the reading. This is how it went.

Starting from memory

The assistant could not find the starting point. Polish civil records are closed for a hundred years for births and eighty for marriages and deaths, so the first two or three generations have to come from the family. It asked for parents and grandparents with maiden names, rough dates and places, and any tree a relative had started online. I phoned my parents and wrote to older relatives, with question lists it drafted in Polish.

Memory was the best source and the least reliable. The maiden name I knew turned out to be spelled two ways in the registers, and a third way in Cyrillic. The breakthrough on my father’s side was not a record at all. The assistant found three graves in a Warsaw cemetery index that could have been my grandfather’s, and I picked the right one, because the stone carried a pet name I knew and an age that fit. Everything behind that grave comes from records. That one link rests on my recognition and a matching age.

Reading the registers

Nearly everything older is public, if you know where to look: Geneteka, which indexes millions of Polish parish entries by name; Metryki and the state archives’ site, with the scans; BaSIA and the Poznań Project for Wielkopolska; the Warsaw Rising Museum’s biographies; Vienna’s city directories; cemetery databases; old newspapers and address books in Polona.

What the assistant did well was read. The books change language as the borders change: Polish until the 1860s, Russian in the Russian partition after 1868, Latin in the eighteenth century, German in Prussian Poznań and in Vienna. One great-grandfather is baptised under the German form of his names in one book and buried under the Polish form in another. The assistant read the scans, checked ages against earlier acts, and kept a rule I lack the patience for: every fact gets a source, with act number, and a grade. Confirmed means it read the scan. Probable means an index row with the right parents and parish. All sixteen great-great-grandparents are there, and 25 of the 32 before them. Then it thins fast, to 2 of 1,024 at generation ten.

It also caught conflicts I would have let through. One birth year would have made a mother sixteen at her first child; the gravestone, read closely, put her birth three years earlier. A death I had been told happened in 1939 turned out, on the stone and in a wartime register, to have been in 1944, in the Uprising.

It was not magic. The biggest index started feeding fake rows to anything that searched it too fast, so we slowed down and checked each answer against an act we already knew. Some sites block scripts or want a signed-in human, so I signed in and ticked the captchas myself. Handwriting beat it too: a father’s name on a death act, a worn last digit on a stone.

Where it stopped

A lot is simply not online. Some 1860s baptism books are still in the parish, not the archive. The books of two parishes east of Warsaw before 1842 are in a diocesan archive. In the Wielkopolska town the baptism index begins in 1700, and everyone at generation nine there was born before that; the earlier books sit in a diocesan reading room. The assistant drafted about two dozen requests, in Polish and German, to parishes, diocesan and state archives, civil registries, cemetery offices and Vienna’s city archive. It could not fill in a captcha or pay a fee, and it sent nothing without my yes. Most of the letters I sent myself, and some still wait on a reply.

Seeing it

The tree is one web page the family can open on a phone, and it is not public. Anyone in the family can put themselves at the centre and see who everyone else is to them, in Polish kinship terms. There is an ancestor table, an hourglass of your family, and a globe where the borders change with the year: the Commonwealth, the partitions, the Duchy of Warsaw, Congress Poland, the Second Republic, the occupation, and now. The first version showed Prussia in 1750 where the Commonwealth should have been; it was redrawn.

What the history looked like

Seen from above, the tree is a slow walk into Warsaw. One line comes from a small town in Wielkopolska where eighteenth-century ancestors were a blacksmith, a furrier, and twice the “proconsul”, which the assistant worked out meant the burgomaster, an office that rotated among the councillors every quarter. After the Prussian takeover of 1793 the books stop giving the title. Another line comes from farming villages east of Warsaw, in Latin books that call the men laboriosi, and reaches the 1760s. A third comes up from the manors south of the city: an estate gardener whose children’s baptisms trace his moves from one manor to the next, and the widower who married his daughter in the 1890s, a sugar-works overseer who had come to his first wedding as a tsarist soldier on leave.

By the early 1900s most of them are in the city: a shoemaker’s journeyman, a servant girl from the north, a fruit merchant, a baker who became a master butcher. A young man born in Prussian Poznań died in a German uniform in Flanders in 1914. A great-grandfather ran a business in Vienna after the First World War, came back to Poland in the 1920s, and was killed in Warsaw in the Uprising. A great-uncle fought with the Polish forces in the West, settled in England and later moved to Italy. The family story about a Romanian princess found no princess, only a great-grandmother born in Czernowitz in the 1890s, a city that changed country while she was in her twenties.

The assistant read faster than I could and kept a source for everything it added; the judgement stayed with my family and me. The tree is honest about the difference, which is the part I am proudest of.