Zakhor

Letter no. 4 · Monday, August 24, 2026

Hello,

Editorial — Zakhor letter Thursday, August 27, 2026

One hundred twenty million Jews

The figure exists, and it possesses an unexpected precision. Sergio DellaPergola, demographer at the Hebrew University, established it in a study published in the journal Genealogy: approximately 120 million Jews have entered the Jewish people, through birth or conversion, from Abraham to today. The calculation accumulates entries over thirty-five centuries; it makes no claim to exactness and its author is the first to say so. But it gives the order of magnitude, and above all it gives a distribution that should command the attention of anyone concerned with memory.

Half of these 120 million were born between 1500 before the common era and 1700. The other half — sixty million people — was born only in the last three centuries.

Fifteen million are living. One hundred five million are dead.

Zakhor has named 505,851 of them. Four out of a thousand.

Why the Bible counts, and why it forbids it

The book of Exodus is not called that in Hebrew. It is called Shemot — Names. The account of a people's liberation opens with a list: "These are the names of the sons of Israel who came to Egypt." Before servitude, before the sea, before Sinai: the names, one by one.

Genesis proceeds in the same manner, by the repeated formula toledot — these are the generations. The first book of Chronicles opens with nine chapters of pure genealogy, from Adam, without a single sentence of narrative. It is the most arid page of the Bible and no one has ever dared to remove it.

And when the Torah counts, it employs an expression that is poorly translated. Se'u et rosh: not "count," but lift up the head of the assembly — and the count is made bemispar shemot, "by the number of names." Biblical census does not add up units: it takes heads and inscribes names.

Which illuminates the prohibition that accompanies it. Counting directly is dangerous — Exodus imposes the half-shekel so as not to number the persons themselves, and the census of David is punished. The Bible counts obsessively and forbids counting. The contradiction is not one: what is forbidden is reducing a person to a number; what is commanded is naming them. Isaiah formulates it definitively — yad vashem, a memorial and a name, "an eternal name that shall not be cut off."

We do nothing else. One hundred twenty million is a statistic. It is one hundred twenty million names that interest us.

What the machine changes — and what it does not

The obstacle has never been conceptual. It has always been arithmetic.

Take a matzevah. One must photograph it, read Hebrew gravings often eroded, transcribe, restore the father's name, convert a date from the Hebrew calendar, normalize the spelling of a surname, distinguish this Moshe ben Yaakov from the eleven others in the same cemetery. For a trained reader: twenty to forty minutes per stone. We possess 456,144 tombs. At thirty minutes per unit, that amounts to approximately 130 person-years at full time — to read what we already have, without recording a single new one.

That is why this work has never been done. Not from indifference: from budgetary impossibility.

A chain of agents now performs the first pass — transcription, date conversion, onomastic matching, proposal for linkage — and leaves to human review what pertains to judgment. The cost per stone drops to a few minutes of validation. One hundred thirty years becomes one. This is the exact point where the shift occurs: the machine knows nothing that we do not know; it makes feasible what was only desirable.

Yad Vashem, which has pursued for seven decades the search for the names of the victims of the Shoah, crossed the threshold in November 2025 of five million names — more than 80 percent of the victims. The institution announces today its recourse to the same methods to find the next 250,000 names. Seven decades, considerable resources, a unique and relatively documented century: this is the standard by which any promise must be calibrated.

We draw from this a rule and we will not negotiate it: the agent proposes archives to cite, never to invent. No fact enters the patrimony without human validation. Each published fact bears its status — established, probable, transmitted, conjectured. A machine error, at this scale, is recopied four hundred thousand times: this is why non-erasure and versioning are not merely moral principles, but technical security devices.

The difficulties

We prefer to display them rather than smooth them over.

Homonymy. Without a stable identifier, surnames repeat infinitely within the same plot. Matching is probabilistic; it must remain revisable.

Absence of sources. For the Maghreb, the Levant, Central Asia, India, the destroyed communities of Europe, the registers no longer exist. Stone is often the only surviving serial source — and it fades.

The temptation of lineage. An identical surname does not imply kinship; a different surname does not exclude it. Onomastics provides direction, it never establishes. It is the discipline most difficult to maintain because it is the one families ask us to break.

Our own undertakings. Deduplication across sources has not been done: our source counter is uniformly at 1 and we cannot yet say that a person is attested by three distinct documents. The memory/history register is not yet encoded at the schema level. The rule of admissibility of notable persons is drafted, not applied.

The asymptote. Most of the 120 million left no trace. We will not name them. But the recent half — sixty million since 1700 — is precisely the one where cemeteries, civil registration records, pinkaissim and lists have survived. That is where the game is played, and it is far from lost.

The satisfactions

They are of a particular nature: they are irreversible. A name restored is not lost again.

A family arrives with three given names and leaves with seven generations. A cemetery in a country where no one visits anymore becomes eight hundred books. A list of subscribers at the end of a Livorno imprint reconstitutes a rabbinical network that the halakhic establishment did not reveal. And a family does not obtain a card: they obtain their book, and they can add to it.

What we need

Material, first. Cemetery photographic campaigns — high resolution, systematic, geolocated, plot by plot: this is acquisition with the highest yield, and the most urgent, because stone fades while we speak of it. Communal registers, ketubot, circumcision records and hevra kaddisha books. Digitizations of manuscripts and printed works, paratexts in priority. Memorial lists. Deposited GEDCOM files. Oral testimonies.

From institutions, we ask first for corpora and scholarly backing, not subsidies. The IIJG for scholarly support, a joint project on communities without archives, and indexing of the Paul Jacobi collection — some 400 studies and 114 typewritten notebooks, incompletely indexed at the National Library of Israel. Consistories and hevrot kaddisha for access to the ground. National and colonial archives for registers. Reciprocity is the principle: what we structure, we return to them indexed.

From families, we ask what no institution possesses. A photograph of a stele taken last month. A corrected reading. A restored Hebrew date. A story transmitted by word of mouth without ever being written. Every contribution declares its register, it is versioned, it is attributed to the one who gave it. We are 1,699 subscribers to this letter: ten names and one photograph per subscriber, that is seventeen thousand people and one thousand seven hundred documents.

Our generation

Of the one hundred twenty million, we are fifteen. One eighth of the total, living at the same time. No generation before ours has represented such a share of the Jewish people of all times.

And none has found itself in this place. We are the last to hear still, in living voice, what Salé, Vilna, Cochin, Mogador or Czernowitz were before the departure. We are the first to have instruments capable of reading four hundred fifty thousand stones. This crossing will not last: the voices fall silent and the stones fade, each at its own pace, all in the same direction.

What is not deposited now will not become an archive later. It will become nothing.

This is why, in this project, the contribution of our generation is decisive — not because it would be more learned or more faithful than those who preceded it, but because it alone stands at the exact point where the last witness and the first tool meet. A photograph of a stele taken this month, a tree deposited, a name read correctly: this is the part that only we can contribute, and no one will take it up after us.

One hundred twenty million names. They depend, today, only on those who are there.

Zakhor — The Great Book of the Jewish people · zakhor.ai

The Grand Book of the Week — Gelman

There are names that cross continents as one crosses a century: by exiling themselves, by reinventing themselves, by carrying with them a part of memory and a part of silence.

19 pages.

→ Read The Grand Book — Gelman

What appeared this week

92 Grand Books were born in seven days. Here they are, by pillar.

Lineages — 26

Oppenheim — 16 p.
Chetboun — 16 p.
Magnus — 16 p.
Morpurgo — 15 p.
Alter — 15 p.
Herzl — 15 p.
and 20 others — see all

Great Figures — 28

Marģers Vestermanis — 17 p.
Meir Wieseltier — 17 p.
Sarah Reisen — 16 p.
Simcha Bunim Shayevitsh — 15 p.
Cyril M. Kornbluth — 15 p.
Joseph ben Solomon — 15 p.
and 22 others — see all

Communities — 8

Themes — 20

Pamięć. Płock 1939 — 14 p.
The Beautiful in Jewish Literature — 10 p.
Judeo-Arab Philosophy, Overview — 10 p.
Deposits of Goods — 9 p.
Chess — 9 p.
Pahlavi Literature, Jews in — 9 p.
and 14 others — see all

Places — 6

Sinagoga Israelita do Recife — 16 p.
Ein Hatzeva Fortress — 14 p.
Old Synagogue of Przemyśl — 12 p.
Kahal Synagogue — 10 p.
Agricultural Colonies in Palestine — 10 p.
Great Choral Synagogue — 9 p.

Institutions — 2

Auto da fé — 9 p.
Council of Four Lands — 9 p.

Texts and Works — 1

Malkhei Rabbanan (Kevod Melakhim) — 9 p.

Four Doors in the Archive

A lineage, a figure, a community, a work. Each week, four others.

The Life of a Lineage — Winkler

At the corner of a street in a city of Central Europe, where commerce mingles with prayer and imperial administration imposes its rules on Jewish communities, the name Winkler takes its definitive form. It is neither a rare nor singular name: Christians and Jews bear it, on both sides of the Rhine and extending to the eastern marches of Galicia.

→ Read — 15 pages

A Great Figure — Gaon of Vilna

There are names that cease to be names to become landmarks. In the firmament of Ashkenazi Judaism, that of Eliyahu ben Shlomo Zalman — whom tradition calls only the Gaon of Vilna, or by his Hebrew acronym the HaGra — shines with such brilliance that an entire religious civilization ultimately came to orient itself according to him.

→ Read — 19 pages

A Community — Sibyl

The name "Sibyl" does not belong, in the strict sense, to a family lineage attested by civil registers, ketubbot, or Sephardic notarial acts. It designates first, in Jewish history, a voice: that of a prophetess whom the Hellenistic diaspora took hold of to speak Greek to the Greeks, while confessing the God of Israel.

→ Read — 7 pages

A Work — Tosefta Bava Batra

The Tosefta Bava Batra — literally "the supplement (or addition) to the Last Gate" — constitutes one of the densest sections of the vast tannaitic compilation known as the Tosefta.

→ Read — 9 pages

Short Stories — The Painted Synagogue

Dura Europos, Euphrates · 244

A collapsed earthen wall preserves for sixteen centuries a prayer hall entirely covered with figures. It was thought such a thing impossible.

Dura Europos is a garrison town perched on a cliff above the Euphrates. In 256, threatened by the Sassanids, its Roman garrison filled the interior of the wall to reinforce it: the buildings lining it were buried under earth.

→ Read the story · All stories

What the nights have found

Our watch turns every night, in thirteen languages. 259 findings this week. Here are three.

An ancient burial cave from the Second Temple period discovered by chance in East Jerusalem

On August 24, 2026, Israeli border guards (Yassam unit) intervening in the Ras al-Amud neighborhood in Silwan (East Jerusalem) as part of an operation related to a violent confrontation between two rival families that left three dead the previous weekend, discovered an ancient burial cave beneath a residential building. The police immediately alerted the Israel Antiquities Authority (IAA), whose teams went to the site for an initial examination. Source

In Saxony-Anhalt, a photogrammetric survey documents the seventy-eight Jewish cemeteries in the region

Since the beginning of 2026, the Gröbzig Synagogue Museum (Museum Synagoge Gröbzig) has been coordinating a comprehensive digital survey of Jewish cemeteries in Saxony-Anhalt — approximately 8,000 gravestones to document, including 2,200 in Magdeburg, 1,300 in Dessau-Roßlau, and 800 in Halle. The project, with a budget close to 200,000 euros, is being conducted by Judaicist Anett Gottschalk, the museum's director, with media technician Philipp Grundmann, who carries out drone photography and 3D modeling of each stone. Source

A century-old Jewish cemetery in Jabalpur vandalized by a land grabber, proceedings still ongoing

The Jewish cemetery in Jabalpur (Madhya Pradesh, India), allocated about a century ago by the British government near the Ranitaal cremation site, is the only Jewish cemetery in Madhya Pradesh and Chhattisgarh — the local community, once numbering 200 families, now has only 4 families. Daniel Ibrahim, a resident of the Civil Lines neighborhood in Jabalpur, observed the degradation of the boundary wall, the disappearance of the gate, and the desecration of graves belonging to three generations of his family. Source

→ All the findings of the week

What Changed on the Site

Cemeteries of the world enter the catalog, country after country

This week was dominated by the addition to the catalog of Jewish cemeteries from around the world — Israel, Egypt, Libya, Algeria, France, Alsace, Russia, Spain, Italy and some twenty other countries, along with their burial registers where they exist. The major project of the unified container for persons has been completed: subscribers, accounts, family trees and private files are now gathered together in a single searchable directory. A wave of large-scale automated access targeted the registration page for several days, without any attempt succeeding, and several protections were put in place in response. The week ended with a thorough overhaul of the home page, tightened and refocused on searching for a loved one.

Monday, August 17 — The cemeteries of Israel, Egypt and Libya enter the catalog
Tuesday, August 18 — The great reorganization of persons begins, and Alsace joins the catalog
Wednesday, August 19 — Each cemetery in the catalog has its location page, and new countries are added
Thursday, August 20 — Search finally finds tombs, and two data leaks are plugged
Friday, August 21 — A massive attack targets the registration page, without success
Saturday, August 22 — A family's book was being misread, and testimonies are now linked to their tomb
Sunday, August 23 — The home page tightens, demographics changes its name

→ The project, day by day

The Week Ahead

Wednesday, August 26 (13 Eloul 5786) — The Ben Ich Haï (Rabbi Yossef Haïm of Baghdad) — yahrzeit · 5669 — 1909
Saturday, August 29 (16 Eloul 5786) — Opening of the first Zionist Congress (Basel) — 1897

What You Have Deposited

3 items this week: Jewish Thinkers of Hispano-Maghrebi Origin: The Encaoua Lineage (Bernard Bensaid), Once Upon a Time in Tunis - book by Tania Bellity (Gad Bellity), Odette Bensaid (Jean Louis Charbit). All 26,051 documents we preserve began with someone who opened a drawer.

1,630 Published Ledgers · 7,390 lineages · 15,440 figures

Those Who Have Joined Us

15 people opened an account this week. We are now 145 keeping this register together.

We do not name them: one comes here to search for one's own, not to be announced.

Join in turn →

What you can do

Write, enrich or simply read a Grand Livre. Each family, each community, each work can have one. If one already exists, you can add what it is missing; if it does not exist, you can bring it into being.

→ Write or enrich a Grand Livre · Browse the Grand Livres

Bear witness about one of your close ones. A grandfather, an aunt, a neighbor from the street before. What you know of a person — where they lived, what they did, what they said — is written nowhere else. The account is signed with your name and attached to the tree.

→ Bear witness about a close one

Import your family tree. If you keep a tree in software — Heredis, Geneanet, MyHeritage, Gramps — it exports to a GEDCOM file that you can deposit here. You remain in control: nothing is published without your action.

→ Import a tree (GEDCOM)

This week, a gesture

Look at what you keep without knowing why. A photo without a caption. A letter in a language you cannot read. A piece of paper at the bottom of a box, that you have never thrown away yet can never explain why.

Deposit an item →

zakhor.ai

Newsletter

The archive is reserved for members