How to research your family line

The order of operations

Almost everyone arrives at this in the wrong order: a DNA test, then a surname list, then an attempt to find documents that confirm what those two suggested. Run it that way and you will find confirmation, because confirmation is always findable. Run it the other way and you find out what is true.

1. 1Build the documented tree back as far as it goesCivil registration, then parish registers — baptism, marriage, burial. Do not skip a generation because the next one is obvious. Most false Sephardic lines enter a tree at exactly the point someone stopped checking.
2. 2Move to notarial and land recordsWills, dowries, property transfers, apprenticeships. These name relationships parish registers omit, and they survive in places where registers burned.
3. 3Search the Inquisition catalogues by person and place, not by surnameRoughly 27,000 Portuguese trials are catalogued. A trial file names parents, grandparents, in-laws, neighbours and occupations — it is often the richest genealogical source that exists for a converso family.
4. 4Look for limpieza de sangre filesPurity-of-blood proceedings were filed to deny converso ancestry, and frequently document it in the course of denying it.
5. 5Only now, test — and test the right thingBy this point you have a specific hypothesis about a specific line. That determines whether Y-DNA, mtDNA or autosomal testing can address it. Without the hypothesis, a test result is a Rorschach blot.

What DNA can actually settle

For any haplogroup question, the answer is one number: the TMRCA of the node where your branch and the nearest Jewish branch actually meet. Everything else on a YFull or FamilyTreeDNA page — flags, tags, badges, notable connections — is decoration hanging off that number.

A shared ancestor at 550 years puts you in the era of expulsions and forced conversions, and is worth pursuing. A shared ancestor at 2,400 BCE tells you your line left the Middle East in the Bronze Age, which is true of a great many European men and settles nothing.

Why an empty match list is not a negative result

In 1348 a mob murdered the Jewish community of Tàrrega and threw the bodies into mass graves. Four of the men have Y-chromosome results. Apply the rule “this clade has no Jewish matches, therefore not Jewish” and you exclude two of them — men buried in a Jewish cemetery, killed for being Jews.

One of them sits in a branch holding three testers worldwide: an Ethiopian, a Saudi and a Belarusian Jew. A clade with three testers carries no information about its composition in either direction. The honest verdict is undetermined — insufficient sampling, and this site says so rather than letting silence read as “no.”

There are four ordinary reasons the databases under-count Jewish lineages: male lines go extinct as a matter of demographic drift; the Inquisition removed them non-randomly, by family; converso descendants test and are recorded as Spanish, Mexican or Filipino rather than Jewish; and Jewish reference labels come overwhelmingly from Ashkenazi testers, because that is who buys the tests. “No Jewish matches” partly measures the market.