Inside the synagogue on Halderstrasse in Augsburg, in the small hours of 10 November 1938, the caretaker was made to unlock his own front door. Twenty or thirty men pushed past him, cut the telephone cable and began pulling the Torah scrolls out of the ark. By dawn the sanctuary was burning, and a fire brigade was working on the roof above them in the deliberate business of putting the fire out.
What a Torah Scroll Is
A Torah scroll — in Hebrew a Sefer Torah, a book of Torah — is a handwritten copy of the five books of Moses, written on parchment and kept in the synagogue’s Torah ark, veiled there by a curtain. Everything about that short description is doing heavy work. The copy is handwritten, not printed; the support is skin, not paper; and the five books travel as a single object. The printed and bound version of the same text that most people recognise as a Bible is called a chumash, and it is a different animal entirely. It may be carried in a bag, read at a table, stacked on a shelf. The scroll may not.
There is a rule in Jewish law — halachah, the body of legal rulings that governs Jewish practice — that the whole Pentateuch must be written out on one scroll. It is a striking requirement, and it is also one whose origins are murky. Scholars working on the En Gedi scroll, a fragment that preserves large stretches of the Dead Sea Scrolls’ biblical text, concluded that at the period it represents there was not yet a rule requiring the entire Torah in a single scroll. As of 2018 no certainty existed about when the rule arose. Which is to say that a form of the book that every synagogue in the world now takes for granted has a date of origin we cannot give.
The scroll is also only the first third of the library. The Hebrew Bible, known as the Tanakh or the Miqra, is arranged in three sections: Torah, the Teaching or Instruction, which is the five books of Moses; Nevi’im, the Prophets; and Ketuvim, the Writings. The set as a whole now runs to twenty-four books, and the form that has become authoritative in Rabbinic Judaism — the Masoretic Text, a medieval reworking of the consonantal text with its vowels, cantillation marks and masoretic annotations — was assembled between the seventh and tenth centuries CE and divided into parashot, the weekly portions or sections, and pesuqim, the verses. A single parashah is what a congregation reads aloud on a given Saturday. Only the first of the three sections is written on skin and rolled. The Prophets and the Writings are normally read from bound codices, which is a different object with a different history, and that difference will matter later.
The skin itself is the first constraint, and it is a constraint of law rather than of aesthetics. Two kinds of animal hide are acceptable for Torah writing. One is gevil, the prepared skin of an animal whose flesh may be eaten and whose sinew was not severed improperly — the full set of rules that govern kosher slaughter. The other is klaf, a smoother and plainer skin treated with salt, flour and m’afatsim, a preparation involving the residue of a wasp enzyme and tree bark. The requirement for skin is not a preference. Paper was available, and paper books were written, but paper was not acceptable for the Torah, and the reason the law gives is durability: the ink must be able to adhere to a surface that flexes, because a scroll is rolled and unrolled constantly, and it must survive that for generations. The ink on a Torah scroll does indeed flake with use, and when too much of a letter has worn away the scroll is pasul, disqualified, and can no longer be used.
Compare the Aleppo Codex, which is in the Shrine of the Book in Jerusalem and which we will come back to. It is a bound manuscript written at Tiberias around 920. Each of its pages is parchment, thirty-three centimetres high by twenty-six and a half wide, written on both sides, and the ink was made from three types of gall, ground and mixed with black soot and iron sulphate. Those are the same raw materials that recur in scroll practice: the gall-nut juice that a scribe may use to treat a tear so that it can be sewn back together, and the soot-and-gall ink that has to sit on skin without sinking into it.
Here is the part that unsettles the modern eye. A Torah scroll has no pages. It has columns — a ruled rectangle, taller than it is wide or about as tall, with writing set in horizontal lines across it — and those columns are arranged in sequence along a single strip of sewn sheets. There is no page number, no recto and verso, no notion of looking up where you were. The reader does not turn to a page; he unrolls until he finds what he needs, and rolls it back afterwards. The physical arrangement is governed by a single elegant constraint that the Rambam set out in the twelfth century: the length of the scroll must equal its circumference when rolled up. Six handbreadths is the ideal length for a scroll on gevil, though on klaf it may vary provided the equality holds. Everything else follows from that. A handbreadth is four thumbbreadths; a thumbbreadth is the width of seven grains of barley laid tightly side by side; a cubit is six handbreadths. Margins are specified in the same units: four thumbbreadths below each column, three above, two between columns. Extra skin is left at the beginning and the end for sewing up and for winding onto the wooden staves.
The writing surface is prepared with a sargel, a ruling tool, and there is an asymmetry in the practice that is easy to miss. The scribe rules the top guide line only. The horizontal lines below it are not drawn; the letters are simply suspended from the top line, each one aligned by eye, each letter resting on an imaginary baseline. Producing that regularity across a sheet of skin, let alone across several hundred thousand letters in sequence, is the whole of the scribe’s difficulty.
And here is where the usual question about page size becomes irrelevant. The classical legal treatment of the scroll’s dimensions is written in lines. The number of lines to a column is fixed by law and by local custom, and it is the operative measurement; the height and width of a column are consequences of the writing, not independent choices. The scribe, in the Rambam’s account, begins by ruling and writing a single sample column to test his own handwriting: seventeen thumbbreadths of writing height, with margins above and below. If too many columns fit across a sheet, he broadens his script; if too few, he narrows it. He repeats the adjustment until the columns align. The text settles the format, not the other way round. Modern scrolls commonly carry forty-two lines to a column, Yemenite scrolls fifty-one, and the Rambam described his own scroll as fifty-one lines across two hundred and twenty-six columns, with the columns for the Song at the Sea and for God’s Song in Deuteronomy left six thumbbreadths wide where the ordinary columns were four.
At either end the sewn strip is attached to two wooden staves, so that it can be rolled. The whole rolling assembly — the staves, the discs that sit between staves and handles, and the handles themselves, two at each end, four in all — is called the atzeret, and the handles are often carved as stylised trees of life, etz chayyim. Over the upper ends sit the crown of Torah, the breastplate and the wimpel, all of which are separate textile and metal objects added when a scroll is dressed for use. The scrolls live inside the ark, and the ark stands behind a parokha, a curtain; the same word, anglicised, gives us parochah, the curtain before the ark that is mentioned in the later literature. In some traditions the scroll is not kept loose in the ark at all but shut into a wooden case called a tik.
Finally, the colophon — the closing note in a manuscript, recording who made it, where, and when. It is where a book tells you about itself, and it is precisely where the record most often goes wrong. The Aleppo Codex was copied in its consonants by a scribe named Shlomo ben Buya’a in Palestine around 920, then verified, vocalised and given masoretic notes by Aaron ben Moses ben Asher, the last and most prominent member of the Ben Asher dynasty of grammarians from Tiberias, whose rivals were the Ben Naphtali school. Ben Asher’s tradition became the accepted one for the Hebrew Bible. But the Aleppo Codex appears not to have been vocalised by Ben Asher himself; a later colophon, added to the manuscript after his death, attributes the vocalisation to him. The Leningrad Codex, of about the same date, has a colophon that claims only that it was corrected from manuscripts written by Ben Asher — there is no evidence he ever saw it. Maimonides, consulting the Aleppo Codex when he wrote the laws of the Torah scroll in his Mishneh Torah, named it simply: the codex used in these works is the codex known in Egypt, which includes twenty-four books, which was in Jerusalem. A book that names itself incorrectly, or is named incorrectly by later hands, is the normal condition of manuscripts, and the colophon is where you find out.
The Labour of the Scribe
Writing such a scroll is a trade, and the trade has a name. The sofer is the professional scribe, and the command to write a Torah scroll is counted in Jewish tradition as the six hundred and thirteenth commandment, appended to the set in the Bible itself. Scribes are contracted, not volunteers, and a well-made scroll of any real size costs, at current prices, somewhere in the tens of thousands of dollars. The work is measured in letters. The Torah contains 304,805 of them, and producing them correctly is not a matter of fluency alone.
The parchment has to be prepared before a word can go on it. Hides are soaked, cleaned, dehaired, stretched, thinned, and the result is ruled with the sargel before the scribe arrives with his reed. A measuring rod of forty to fifty thumbbreadths, divided into halves and quarters, keeps the columns consistent; a silk cord is used for the finer measurements. On the prepared skin the scribe writes his sample column, measures it, and adjusts his script until the columns come out even across the whole sheet. This is why the line count is a real constraint rather than a convention. Once the number of lines is settled, the sheet length, the margins and the number of columns that will fit on each skin are all determined, and the columns per skin are bounded by law: somewhere between three and eight. If a skin would accommodate nine, it is divided. At the very end of the scroll a single column holding a single verse is allowed to stand alone.
The sheets are then sewn together. The thread is sinew — in the legal sense, sinew from an animal of a species permitted for food — and the requirement to sew the skins is itself one of the commandments given at Sinai. The sewing is not run the full height of the sheet, because a seam carried edge to edge tears the skin; space is left at the top and the bottom. Rambam is explicit about the arithmetic of damage: a tear running across two lines may be sewn back, a tear across three or more may not, unless the sheet has been treated with gall-nut juice. When the strip is complete the two wooden staves are attached, one at each end, and the object is a scroll.
What makes the labour worth dwelling on is the failure mode. A slip is corrected by scraping the letters off the skin and writing them again. That is ordinary, and every scribe does it. The exceptional case is different. If it is a divine name that has been written wrongly, the page cannot simply be corrected: it must be cut out and a new one sewn in its place, and the replacement must be copied from the beginning again, in sequence, so that the scroll remains a single unbroken copy. The old page is treated with respect and either buried or placed in a genizah, a repository for discarded sacred writing — never burned, never thrown away. This is the point at which the object outruns its own tolerances. There is no lawful way to destroy a Torah scroll. You can damage it past use and you can then bury it, but deliberately burning one is not an option the law provides for, because burning sacred writing is exactly the sin. To destroy a scroll correctly you have to commit a different wrong.
Getting to the end takes a long time. Standard accounts put the writing alone at something like a year and a half, and that is before the preparation of the skins and before the repair time of the scribe. The work is judged by people who know the Masoretic text letter by letter, because an error that is invisible to a reader is still an error. When the last column is finished there is a celebration, the hachnosas sefer Torah, the receiving of the scroll: the community sings, and the men of the congregation dance with it in their arms. Reading from the scroll is done with a yad, a pointer, held in the hand so that the parchment is not touched, and the reader wears a tallit, the prayer shawl, with its edge drawn across the scroll.
This is the thing that makes a Torah scroll unlike any other old book. It is not a carrier of text. It is an execution of a text, and the execution has a date, a place, a hand and a set of mistakes. Some of those mistakes are recoverable, in principle, if a scribe made them and a record survives. Almost none of them survive.
The Synagogue at Halderstrasse
The building the scrolls were kept in was new enough that the town had almost no time to get used to it. For most of the modern period before the nineteenth century, Jews were not permitted to live in the city of Augsburg at all: they had been expelled in 1438, were made to sew yellow rings onto their clothing, and were not readmitted until 1803. Those who did live in the city afterwards still paid an annual levy for the privilege of entering through the Gögginger gate at the Königsplatz, and they had to leave again before dark. The first synagogue was built around 1800 not in the city but in the district of Kriegshaber; a second, on the Wintergasse, followed in 1858. The city’s Jewish population grew sharply in the last decades of the nineteenth century as people moved in from the villages of the surrounding countryside, and the Wintergasse building, repeatedly altered and extended, simply stopped being big enough. A synagogue building association was founded in 1896 and bought the former Degmair garden property on Halderstrasse in 1903.
The competition was won by Fritz Landauer, born 1883, and Heinrich Lömpel, born 1877 — two architects whose names are unusually well preserved, and who between them designed a building that historians still read as a turning point. The foundation stone was laid on 30 April 1914. The war intervened, construction ran on, and the synagogue was formally inaugurated on 4 April 1917. The Jewish Museum Augsburg Swabien dates the construction rather more broadly as 1913 to 1917; the two accounts differ only in where you start counting.
What they built is a Greek cross in plan, with a reinforced concrete dome over the crossing and four short gable roofs over the arms of the cross, flanked by four stone lions. The sanctuary faces east. It does not stand on the street but is set back behind two symmetrically arranged community buildings, with an entrance hall of three arches between them and lateral wings connecting the whole composition. The east wing holds a wedding hall and a weekday prayer room; the west wing holds a vestibule, a cloakroom and a fountain for ritual handwashing; and the planted courtyard at the centre has a sphinx fountain in it. The decoration mixes Art Nouveau ornament with Byzantine and Orientalizing detail, and the effect has been read both as the architectural ambitions of a confident community and as an expression of a new approach to synagogue design associated with Reform Judaism. The synagogue and its community buildings are protected architectural monuments.
Its organ was a serious instrument. Koulen and Sohn of Augsburg installed it in 1917: two manuals, thirty-two stops, pneumatic action, for something like fourteen thousand gold marks. The congregation numbered roughly twelve hundred members before the war. The community had civic equality in the city for two generations by the time it moved in, and the building is best understood as what that felt like to write down.
Then everything changed in a single night.
The Night of 9 to 10 November 1938
The account that has held the firmest is short and comes from the Schwäbische Landeszeitung of 7 November 1958, twenty years after the event, and it is corroborated closely by the city’s own memorial account. In the early hours of the night of 9 to 10 November, around twenty or thirty men broke into the synagogue on Halderstrasse and threatened the caretaker with a pistol when he opened the door. The telephone lines were cut, so that no one could call for help. The interior was wrecked and the sanctuary was set fire to.
The City of Augsburg’s own account of the pogrom night names what was destroyed, and the list is worth reading slowly because it is a list of a congregation’s entire material existence. Gone were the glass prism chandeliers. Gone were the Torah scrolls with their precious silver ornament. Gone were the Torah ark curtains, the parokhot, and other ritual objects. Gone, too, were the community’s own records — the documents, the registers, the paperwork by which a congregation knows who its members are.
And then comes the sentence that makes this case unlike the others. The synagogue building itself was preserved from total destruction only because the fire brigade extinguished the fire — not out of any regard for the building, but to protect the surrounding structures and, above all, a neighbouring fuel depot. A Nazi fire crew, working at speed, in the middle of the night, saved the shell of the building by stopping the fire from reaching the oil.
It is worth sitting with how strange that is. In the ordinary course of the pogrom, a synagogue that burned was a synagogue that was destroyed, because fire did the work that demolition crews would otherwise have had to do slowly and expensively. Here the fire started and then was deliberately stopped, and the result was the worst of the available outcomes: the building went on standing while everything inside it went away. The dome stayed on the four lions. The Greek cross stayed on its plan. The room stayed a room. The scrolls did not.
The regime then inverted the story. Gestapo officers inspected the site and alleged that members of the Jewish community had started the fire themselves — a charge that has the merit of being exactly the sort of thing one says when the truth is uncomfortable, and that the extinguishing of the blaze by German firemen to protect a German fuel depot makes very hard to sustain. The city’s account goes on to describe the rest of the night with the flatness that such records usually achieve: the damage had to be repaired at the victims’ own expense, and a punitive Sühneleistung — an atonement payment — was imposed on the survivors as well. The Jewish men were held in the municipal jail at the Katzenstadel. After about eight days’ detention, some one hundred and fifty of them were taken to Dachau, where they were kept for at least a month. Within the following three months, eighty-nine Augsburg Jews left the city as emigrants. Leopold Rieser, a lawyer, was the first Augsburg Jew murdered.
There is a detail that survives only in local memory and that I flag as such, because I have found it in the account kept by the Augsburg branch of the association of former victims of the Nazi regime and in no documentary source: after the fire, Jewish men in coats and top hats were made to load the desecrated Torah scrolls onto lorries. Whatever the provenance of that detail, it describes precisely the situation the synagogue was left in — men in formal dress, in daylight, carrying out of their own building the objects that had made the building a synagogue at all.
Afterwards the complex did not stand empty, which is its own kind of injury. The sanctuary was unusable for worship. Municipal offices, including welfare and clothing distribution services, moved into the community buildings facing the street. The complex became an assembly point for the deportation of Jewish families from Bavarian Swabia. From 1941 until the end of the war the municipal theatre used the sanctuary to store scenery. From 1942 the community buildings were occupied by the National Socialist People’s Welfare organisation and a regional Luftwaffe staff, and an anti-aircraft gun position was installed on top of the dome. The building was put to every use except the one for which it had been raised.
One object did leave the building intact, in a way nobody had planned. The organ survived the fire but not in any playable condition, and it was clear that the Nazis would strip it for its metal. According to the account in the English-language record of the synagogue, the rabbi sold it in 1940 to a priest he knew in Weßling, in order to put it beyond reach; in 1941 it was installed in the Catholic parish church of Christ the King there, where it was too large for the building, so parts of the ceiling were cut away to admit its largest pipes, which were subsequently bent to fit. It was restored, and was still in use there in 2017. Other accounts of the same organ say that the National Socialist party was the seller. The two versions disagree about who profited and who acted, and the record does not settle between them.
What Cannot Be Rebuilt
It is easy to say that the synagogue was saved and the scrolls were destroyed, as though both were the same kind of thing and one simply had better luck. They are not the same kind of thing, and the difference is the whole of the subject.
A building is reproducible. Landauer and Lömpel’s plans could have been put up again tomorrow by an architect with the drawings, and in the 1970s and 1980s they very nearly were, in the sense described shortly. Concrete can be recast, stone can be cut, a dome can be rebuilt over the same crossing plan. What gets restored is the design, and the design is separable from any particular instance of it. A Torah scroll is not like that. A scroll is the writing of one hand at one moment, made for one congregation, out of skins that came from particular animals, with a particular set of errors in it. There is no master copy from which a second Augsburg scroll could be made. The word the tradition uses for the Torah as the object of the commandment is not book but writing, and writing does not have a second copy. It has copies made from a copy, which is a different thing, and the difference is the subject of the first half of this essay.
The law makes the asymmetry even sharper, and this is the part that is genuinely difficult for a modern reader. Because a divine name written in error cannot be corrected on the page, it must be cut out and replaced, and the discarded page must be buried or placed in a genizah and never burned or discarded. There is consequently no procedure by which a Torah scroll may lawfully be destroyed. It can be allowed to become unfit; it cannot then be thrown away. To burn one deliberately is not a permitted disposal but a fresh offence against the object, which means that the destruction of a scroll is the one act for which the scroll’s own rules provide no remedy. In halachah, the term halitzah — the formal release that dissolves a marriage without the ordinary process — belongs to family law and has no application here at all, which is precisely the point: the law has no word for this, because the law did not suppose it would come up.
Compare what a community does when it loses the manuscript but not the room, or loses the manuscript at all. For six hundred years the Aleppo community kept the codex in the Central Synagogue of Aleppo, together with three other biblical manuscripts, in a special cupboard — later an iron safe — in a basement chapel that was supposed to be the Cave of Elijah. It was regarded as the community’s most sacred possession. People in trouble came and prayed before it, and oaths were taken by it. Jews from all over the world sent queries asking whether particular textual details should be checked against it, and the answers were preserved in the responsa literature. In 1943, Umberto Cassuto was the only modern scholar permitted to compare it with a printed Hebrew Bible and take notes of the differences. The community guarded it so closely that when Paul Kahle tried in the 1920s to obtain a photographic copy for a new edition of the Bibliia Hebraica, he failed, and was obliged to use the Leningrad Codex instead. Its protection was a form of its unavailability, and it cost scholarship real access.
In 1947, rioters enraged by the United Nations partition plan for Palestine burned down the synagogue in Aleppo. The codex disappeared. It resurfaced in Israel in 1958, smuggled out by the Syrian Jew Murad Faham and presented to the president of the state, Yitzhak Ben-Zvi — and when it was examined, roughly forty per cent of the manuscript was gone, including the majority of the Torah section. Of an estimated four hundred and eighty-seven pages, two hundred and ninety-four survived. Only two further leaves have been recovered since, and no more than a handful of pages. It was long assumed that the missing pages burned in the synagogue; analysis showed no evidence that fire ever reached the codex, and the dark marks on the pages turned out to be fungus. Witness accounts from Aleppo describing the codex shortly after the fire consistently reported it complete or nearly complete, and then there is no account of it at all for more than a decade. The community sued for its return in 1958 and lost, and the proceedings were suppressed.
Set the two next to each other. In Aleppo the room burned and most of the book survived. In Augsburg the room survived and the book did not. Both outcomes were accidents, and they were accidents of the same order — not of a policy, not of an ideology about manuscripts, but of where a fire happened to spread and where it happened to stop.
What Is Left
The building took further damage from the Allied air raids on Augsburg, and the interior remained destroyed for years afterwards. Provisional repairs allowed the synagogue to reopen on a limited basis by 16 May 1945. The small prayer room in the complex, known as the Kleine Synagoge, returned to communal use on 15 December 1963. The main restoration was carried out between 1974 and 1985 under Julius Spokojny, who had spent a long time as chairman of the Israelitische Kultusgemeinde Schwaben-Augsburg. The restored sanctuary was formally rededicated on 1 September 1985, and the same year the Jewish Museum Augsburg Swabia opened in the complex. The synagogue is visited as part of a museum visit.
So a synagogue in Augsburg is a room again, in the strongest and most literal sense: it is used, on Fridays and Saturdays, by a congregation that worships in it. Each 9 November the City of Augsburg and the community hold a commemoration hour in the synagogue together. What it is not is a room that holds its scrolls. Nobody has been able to say how many there were. The city’s account of the losses lists the Torah scrolls with their silver ornament without giving a number, and neither the museum, the city lexicon, the crime atlas nor the synagogue’s own history supplies one. The building that survives is the part for which drawings exist.
This is the wider point, and it is not a sentimental one. Manuscript survival is not a measure of importance. It is a measure of accident. The Aleppo Codex is the most authoritative text we have for the Masoretic tradition, and it is authoritative largely because it sat for six hundred years in one cupboard in one cellar in one building, was guarded jealously, survived a riot that destroyed the building around it, was smuggled across a frontier by one man, and happened to resurface in a place where scholars could examine it. Every one of those steps was luck dressed as procedure. The scrolls of Augsburg are not a manuscript of anything, because nobody recorded which scribe wrote them, which congregation commissioned them, in which year, from which animals’ skins, or what mistakes they contained. That information was in the room when the room was emptied. It went out on lorries or into the fire along with everything else, and the record is now silent on all of it.
What remains of the interior is a photograph from around 1945, held by the Jewish Museum Augsburg Schwaben, showing a demolished sanctuary. The most durable thing the fire brigade saved in Augsburg that night was the shape of the loss.
References
- Pogromnacht 1938 in Augsburg — Stadt Augsburg, Erinnerungskultur, updated 15 June 2026
- Reichspogromnacht — Wissner Stadtlexikon Augsburg, Wissner Verlag
- November pogroms of 1938 in Augsburg — NaziCrimesAtlas, citing Schwäbische Landeszeitung, 7 November 1958
- Augsburg Synagogue — Grokipedia
- Die Augsburger Synagogen — Jüdisches Museum Augsburg Schwaben
- Synagogen, jüdische Friedhöfe und Gedenkorte in Augsburg — Vereinigung der Verfolgten des Naziregimes, Kreisvereinigung Augsburg, page dated 3 April 2022
- Torah scroll — Grokipedia
- Hebrew Bible — Grokipedia
- Aleppo Codex — Grokipedia
- Maimonides, Hilkhot Sefer Torah, chapter 9: the physical dimensions, proportions and sewing of a Torah scroll — English rendering by Rabbi Yaakov Goldstein, Shulchan Aruch HaRav, 29 April 2026