What Cannot Be True

A contradiction in the record — 29

The dates as written cannot both be true. This is the only kind that is really a finding about the past rather than about the tree.

A wrong link — 4

Too far apart for any slop in the dates. The join is wrong, not the dates — usually a reused name, or two people merged.

A generation skipped — 10

The parent is barely older than the child, because a grandparent has been recorded as a parent.

A sex and a name that disagree — 26

The recorded sex contradicts a given name that this archive's own records show is overwhelmingly one sex. Which of the two fields is wrong is not decided here.

Not an error at all — 17

A hedged date (BEF, AFT, Maybe) sitting against a firm one. The hedge is doing its job.

The clearest case is not on the direct line at all

Vladimira Baricević, born 12 December 1907, is entered as the mother of ten Boras children born 1888 to 1911 — nine of them before she was born, one of them in the same month. She is not a doubtful reading or a loose date. She is somebody else's daughter married into the wrong generation, and every child hangs off her.

Which is exactly what happened to Maria Catharina Orešković — four children born when she was two, five and six, plus the son born the year before her. Question 4d worked that one out the slow way, from three parish registers. This check finds the same shape in seconds, without leaving the file.

Thirty-eight of the 113 sit on people with pages here. The rest are relatives who appear in someone's family list without a page of their own — still worth naming, because a wrong join is visible from both ends.

The errors were voting on whether they were errors

The case that prompted this is Ivana Blažević, recorded M. Ivana is borne here by 21 women and 3 people recorded M. Counted whole, that is 87 per cent — under any threshold worth setting, and the check says nothing. The three wrong records were holding the name below the line that would have caught them.

So the majority is counted without the person being judged. Set Ivana Blažević aside and the other 23 are 91 per cent women, and all three are found. The same thing happens one group over: the four “women” in the Franjo tally are not other names caught by the fold — they are the four records this check reports.

Two couples are recorded as each other

Lucija Lončarić Papić is recorded M and entered as the husband of a wife called Ivan, who is recorded F. All 124 people here bearing a form of Ivan are men; 50 of the 54 bearing a form of Lucija are women. Lucia Žubrinić and Michaël Kostelacz are the same shape again. In both, the sex fields follow the place in the family and not the name — which is one swap, recorded twice, and visible only because both halves were checked.

Half of each couple has no page here, being named only in the other's family list. This archive has three times mistaken an unpublished person for no person at all. Both halves are reported.

One letter, and the tree wrote it itself

Joannam Papić, born 1845 at Selce, is recorded M. Joannam is the accusative of Joanna, a woman. The accusative of the man's name Joannes is Joannem, with an e — which is why the archive's given-name table files joannam under Ivana and joannem under Ivan. The finding is carried entirely by a letter somebody copied out of a register, and the same page that explains why the case ending is there is the page that settles which name it belongs to.

What it still misses, and the bug it found on the way

Counting the bearers exposed a fault in the counting. Eighteen name forms are filed under two or three languages — Maria is Latin, Italian and Hungarian — and the tally added a person once per language. The one person written Maria and recorded M was three men, and Mihovil had 87 bearers where it has 39. Those numbers were printed on the name page until it moved to a shared component; they now feed this check instead, which is a worse place for a count to be wrong. Fixed at the source, and the twenty-two findings did not change — which is the only reassuring thing about it.

And the check is deliberately deaf in one ear. The tallies count published people, so someone without a page contributes nothing to set aside and is judged by the whole group, wrong records included. Ivana Juhas, recorded M, is that person, and this check does not find her. It errs towards silence, which is the right direction for something that stops a deploy — but the gap is written down rather than left in the gaps.

1. Three parents die in 1845 and have children in 1847 — and 1845 is a guess

Elias Žubrinić (d. 1845, child born 1847), Phillipus Žubrinić (d. 1845, child born 1847) and a Maria (d. 1845, child born 1847). Three of the same shape in one family in one year.

That looked like an epidemic. This archive has suspected one since question 4b, because Elias Žubrinić and his wife Ana Untersinger both died in 1845 leaving seven children. Tested, it is not an epidemic.

The reason is in how the dates are written. Six people in this family carry the bare string 1845 — no day, no month, no place, no source: Elias Žubrinić, Ana Untersinger, Phillipus Žubrinić, Martinus Žubrinić, a Maria born 1838 and a Maria born 1809, plus Nicolaus Kraljić. The only two 1845 deaths in this archive that carry a real date — Jelena Katnić on 21 November and Mihovil Kalanj on 13 March — are not in this group at all.

An epidemic recorded in a death register produces days and months. A year applied in bulk produces exactly this. So “1845” here is far better read as somebody's estimate for a group of people who dropped out of the record together — and if it is an estimate, the children born in 1847 are not impossible, merely later than a guess.

This weakens question 4b rather than answering it, and the request to Gospić should now ask a different thing: not “was there an epidemic in 1845” but “when did these six actually die”. A single decade of the Otočac death register would still settle it.

2. The Boras generation, skipped

Eleven rows across four people, and they are one family making the same mistake at both ends of itself: a generation has been skipped twice.

At the bottom. Budimir Boras (1894–1954) and Marica Milković (b. 1904) have ten children. Eight of them — 1923, 1924, 1924, 1926, 1927, 1929, 1930, 1933 — are a coherent set, born when Marica was between nineteen and twenty-nine. The other two are not: Domagoj in 1946 (Marica at 42) and Vladimira in 1960 (Marica at 56, and six years after Budimir died). Those two are almost certainly grandchildren — children of one of the eight — attached to the grandparents instead.

At the top, the same thing. Joannes Ivan Boras, born 1855, is given Vladimira Baricević, born 1907, as his wife, and fourteen children hang off her — nine of them born before she was. A woman born in 1907 is the right age to have married one of his sons, and that is the likeliest reading: she has been put in the wrong generation, and Joannes's actual wife is missing from the record entirely.

One woman in the wrong slot accounts for eleven of the sixty-three rows on this page — which is the whole argument for counting causes rather than symptoms.

3. One typing slip that produced three impossibilities

Georgius Prpić, entered as born 26 March 1962 to parents who died in 1888 and 1891 — reported here as a child born after his father's death, a child born after his mother's, and a mother of 139. His own birthplace settles it: Krivi Put 136, Senj, Lika, Austria-Hungary, a state that ceased to exist in 1918, at his parents' own house. Everything agrees with 1862. What changed →

4. Two that are already open questions

Maria Catharina Orešković, born 1835, as mother of a boy baptised 1834 — this is question 4d, worked out the slow way from three parish registers. And Michael “Mate” Žubrinić, dead in 1863 with a son born in 1880, which is the one man or two question the Gospić letter exists to settle.

5. And six single slips with nothing behind them