Corrections

was “The file for marriage act 26 (Falco × Crisci) is not in the 1814 processetti. Do not search these two again.”

now It was there — bundle 16, images 265–274 — and it is the richest document this archive has found.

how An earlier session worked the volume from the back, found a publication act numbered 25 dated 20 November 1814 and then the 1815 divider, and concluded the volume stopped short of act 26. But processetti BUNDLES are numbered in their own sequence, in the order the banns were published — not by the number the marriage later received in the marriage register. The 1814 volume holds SEVENTEEN bundles; Vincenzo and Adriana published in September and married in December, so their file is bundle 16 while their marriage is act 26. The «25» that stopped the earlier search was a publication act number inside bundle 17, the last of the year.

This is the most expensive kind of error an archive can make — a wrong negative, written down as a rule, that stops anyone looking again. It stayed in QUEUE.md under “Dead ends, recorded so they are not repeated” for weeks. The standing rule is now: in a processetti volume match on names and publication date, never on the marriage act number — and every “do not search again” note must carry the reasoning that produced it, so the reasoning can be checked.

Caught by re-reading the volume. · Where it lives now → · the plate

was Vincenzo Falco, generation three, born 1793

now Baptised 1 February 1792 — and christened Vincenzo RAFFAELE

how His own baptism extract, copied from the parish register (liber IX, fol. 47v) by the archpriest Giacinto Morgillo on 13 September 1814 and filed in his marriage dossier.

1793 came from his marriage act's stated age of 21, which was itself wrong. The date is corrected, and the old one is left standing here because the seven-way age conflict it caused is one of the most-cited things on this site. Knowing the real date does not erase the conflict — it explains it.

Caught by the 1814 processetti. · Where it lives now → · the plate

was Generation three's wife published two ways — Andreana CRISCI, and Andriana CIOFFI from the 1839 death register

now CRISCI. Baptised ADRIANA on 30 November 1796, daughter of Arcangelo Crisci and Sara Ruggiero.

how Six documents of 1814 — the dossier cover sheet, both notarial consent acts, both civil banns, and her own baptism extract — all say Crisci, and four of them name her parents.

The 1839 reading was published as a live conflict rather than suppressed, and it stays visible here. It was either the clerk's error or mine. The baptism also explains the variants: Adriana at the font becomes Andreana, Andriana and Andrea in the registers. They are one woman.

Caught by the 1814 processetti. · Where it lives now → · the plate

was The Lines diagram drew the ZAMPIELLO as a separate stream running alongside this family, omitted SAN FELICE A CANCELLO and FORCHIA as places entirely, and used one visual grammar — a dot on a lane — for places, for people, and once for a ship.

now A place-spine. The Zampiello, Annecchino and Morgillo are drawn JOINING the line at the generation and town where they join. Names sit under the place they belong to.

how Checked the diagram's places against places.json and the atlas. It showed seven places; it was missing San Felice a Cancello — where Angela Rosa Morgillo, generation four's wife, was born — and never drew Forchia as the Annecchino's town even though that marriage is the reason generation eight was born there. Two of the five lanes had people, not places, as their nodes, and one had the ship Sydney.

The Zampiello error was the worst of it: drawn as a parallel river, they read as a different family being compared with this one, when Francesca Zampiello married Carmine Falco in 1955 and from generation eight onward they ARE this family. A diagram that is wrong is worse than no diagram, because nobody checks a picture against the prose. Raised by David, who spotted all three faults from the image alone.

Caught by the reader. · Where it lives now →

was “San Felice a Cancello is a separate comune with its own registers, NONE OF WHICH THIS ARCHIVE HAS READ.” San Felice is where generation four's wife, Angela Rosa Morgillo, was born.

now Its registers are on Antenati and have been all along — Archivio di Stato di Caserta, marriages from 1809 at an_ua50221. The 1817 marriages have now been read.

how A search endpoint that returns ark ids in plain HTML: search-registry/?localita=<town>. Resolving those arks gives the series arithmetic for any town. The archive's own survey covers five Valle Caudina towns and does not include San Felice at all.

The wording said “has not read” and was heard, including by me, as “cannot read”. In fact nobody had looked. It was missed because Antenati catalogues the town under its older name, «San Felice (oggi San Felice a Cancello)», and because this archive reached Antenati by guessing arks from an arithmetic rule derived by hand for Arienzo alone — a method that can only ever find what it already knows about.

Caught by looking properly. · Where it lives now →

was The rebuilt Lines diagram printed generation nine's FULL DATE OF BIRTH beside her name, under the heading BORN HERE — on /lines and, because the home page embeds the same figure, on the front page of the site. She is living.

now «9 · Giuseppina Falco — living — named only», under the heading LIVING. The 1955 on the Benevento stop now belongs to generation eight's marriage, where it is a fact about two people who are dead.

how Sweeping the built site for the date and finding it in the raw HTML of two pages. The visible text of /direct-line was correct; the diagram was not.

This archive has now leaked this exact birth year THREE times: once through a redaction function reused from the wrong place, once through the timeline's geometry — and now through a diagram I built myself on the day this correction was written, hours after re-reading the standing rule. Textual fixes have not held, so THE FIX IS STRUCTURAL: the diagram now looks every date up against the person graph and prints «living — named only» instead, whatever the underlying data says. Data can be wrong; the guard cannot be bypassed by editing a string. And this entry had to be written twice: the first version QUOTED THE DATE IT WAS CORRECTING, which republished it on a page nobody would think to check. The release gate caught that on its first run. A correction must describe a leak without reproducing it.

Caught by auditing the built HTML, not the source. · Where it lives now →

was «No index, no indice decennale, no marginal surnames — all three checked and absent. ~90 acts/year, ~1,350 total. ONE SCREENSHOT PER ACT.» That recipe governed twenty-eight years of death-register sweeps.

now Antenati lists 100 INDEX VOLUMES for Arienzo alone — 39 for deaths, 44 for marriages, 17 for «diversi» — and 79 more for San Felice a Cancello. Arienzo death indexes run 1809 = an_ua14271, +1 per year, verified at 1818, 1828 and 1834. Each is ONE TO FOUR IMAGES and lists every death of the year with BOTH PARENTS NAMED.

how Reading the catalogue's own facet list instead of paging the volumes. The facet is spelled «Morti, indice» and «Matrimoni, processetti» — query for «Processetti» alone and you get nothing, which is part of how this was missed.

This is the third wrong negative of the same shape in a single day — the 1814 processetto «not in the volume», San Felice «unread», and now indexes «absent». Each came from checking once, concluding, and writing the conclusion down as a rule that stopped anyone looking again. The acts are still worth reading for what an index omits — street, hour, declarants, the «parente» wording this archive's best finds turn on — but the register is no longer the place to START. Two cautions are published with it: the sort order changes year to year, and the scans are not always complete, so an absence in an index is not a negative.

Caught by reading the catalogue. · Where it lives now →

was Chiara Rivetti died on 3 SEPTEMBER 1843. Published on /rivetti, /direct-line and in the memory of this project, and used to date the widowhood that brackets Pasquale Falco's death.

now 13 SEPTEMBER 1843. The act says «il dì TREDICI del mese di Settembre» in its heading and «nel giorno TREDICI … alle ore venti» in the declaration — twice. TREDICI was read as TRE.

how The 1843 death INDEX, found the same day, gives «Rivetti Chiara … 13 7bre 1843». The act was then re-read at native resolution and says thirteen in both places.

It also removes an oddity nobody had questioned: the parish buried her on 16 September, which is a strange thirteen-day wait after a death on the 3rd and an ordinary three-day one after the 13th. THIS IS THE INDEX METHOD EARNING ITS KEEP ON ITS FIRST CONTROL TEST — it was run against a known answer precisely to see whether it could be trusted, and it caught an error in the known answer.

Caught by the 1843 death index. · Where it lives now → · the plate

was «GIOACCHINO FALCO is a strong candidate for a fifth child of Matteo and Francesca, THOUGH NO RECORD YET NAMES HIS PARENTS.» Carried that way for years, and he sat among the people this archive could join to nobody.

now «7. FALCO GIOACCHINO | Arienzo | CONTADINO | FALCO MATTEO, e CRISCI FRANCESCA | 14 Febrajo» — Arienzo deaths 1851. He is a documented son of the founding couple, who now have FIVE named children.

how The alphabetical index bound at the front of the 1851 death register — four images. Not a register read act by act.

Recorded because the archive was RIGHT and refused to say so without a record. The suspicion rested on an inference from 1816, when Gioacchino and Giuseppe Falco declared the death of Crescenzo Rivetti as «parenti del defunto», Crescenzo being married to their sister Angela. Holding that as a candidate rather than a fact for years was the correct call, and this is what it looks like when the discipline pays.

Caught by the 1851 death index. · Where it lives now → · the plate

was Matteo Falco dead by 17 October 1822 — the earliest act that calls him defunto

now Dead by 8 December 1814

how A notarial act of that date names «Pasquale Falco del FU Matteo». Nothing published was wrong; the window simply closes almost eight years earlier.

Recorded here because this site publishes the narrowing of a claim as readily as its overturning. The bracket on Matteo's death is now [1786, 8 Dec 1814].

Caught by the 1814 processetti. · Where it lives now → · the plate

was This page said, hours earlier on the same day: “Bundle 25's publication is dated 20 November, two months later than ours.”

now There is no bundle 25. The 1814 volume holds seventeen bundles. The «25» was a publication act number inside bundle 17.

how Sweeping all 290 images of the volume as contact sheets and reading every bundle cover. Each is headed «Numero N — Comune Corpo d'Arienzo»; they run 1 to 17, and bundle 17 is Francesco Crisci fu Alessandro × Teresa Morgillo, published 20 November 1814.

I repeated a number out of the earlier session's note while writing the correction that overturned that very note — without checking it. The mechanism I described was right and is unchanged; the illustration was wrong. It is published because a correction that quietly acquires its own errors is worse than no correction, and because this is exactly the failure the same day's other entry warns about: a figure inherited rather than verified.

Caught by reading all seventeen bundle covers. · Where it lives now →

was Andreana Crisci “married at about sixteen” — inferred from her stated age of twenty in her son's 1818 birth act, which implied a birth about 1798.

now She was born 30 November 1796 and married at EIGHTEEN. The 1818 act understates her by about a year and a half.

how Her baptism extract in the 1814 marriage dossier. The 1818 birth act was then re-read at native resolution: it does say «venti», twenty, and it does say «ANDREANA CRISCI» — the surname is unambiguous, which settles the 1839 «Cioffi» reading a second way.

“Married at sixteen” is the kind of detail a reader remembers and repeats, and it was this archive's own inference from a clerk's guess rather than anything a record said. Eighteen is still young. The same act gets her husband's age right — twenty-seven, when he was twenty-six and three months, which is correct by the year-a-man-is-in reckoning these clerks used. Right about him, wrong about her, in one sentence.

Caught by the 1814 processetti. · Where it lives now → · the plate

was The apothecary's mother recorded as «LUCIA COPPA(?)/CIOFFI» — read at 3400px and left with a question mark and an alternative.

now LUCIA CIOFFI. Read at native resolution: «e di Lucia Cioffi Defunta».

how Re-fetching the act as a IIIF region at the image's own resolution rather than scaling the whole leaf. The double-f ligature is unmistakable at full size and was not at 3400px.

A name carried with a question mark for weeks is a name nobody can search for. It also changes what the record means: CIOFFI is a gentry surname at Arienzo, and in February 1815 a Don Saverio CIOFFI was married to a Donna Anna Maria RIVETTI. The two families married in both directions.

Caught by re-reading at native resolution. · Where it lives now →

was Orazio Morgillo × Vincenza Pesce would be found in the San Felice marriage registers of 1816 or 1817.

now Neither year holds them. Both volumes have been read act by act and the couple is not in either.

how San Felice a Cancello, Matrimoni 1817 (an_ua50229, 22 images) and 1816 (an_ua50228, 26 images), swept in full.

Published as a negative so the two years are not read again. The couple was reached anyway, and by a better route: their daughter's BIRTH act, which names both parents outright. The lesson is recorded on /morgillo — when the target is a couple and a child's date is known, go to the birth register, because a birth act names both parents and a marriage act has to be found first. The marriage is still open; the remaining window is San Felice 1813-1815, or another parish if Vincenza was not of the town.

Caught by reading both volumes. · Where it lives now →

was «Almost certainly the same woman, married in 1814 and buried before the winter was out. Published as PROBABLE, not proved.»

now Established. Bundle 13 carries Margarita Crisci's own baptism extract — «nata ex ANGELO CRISCI et GRAZIA, 27 September 1780» — and the death extracts of both those parents. The 1815 death act names the same two parents, both dead. It is the same woman.

how Reading inside the bundle rather than only its cover. A processetto had to carry death certificates for any dead parent or grandparent, so it also yielded her grandfather: THOMAS CRISCI, widower, dead 21 March 1789 aged forty-six.

Recorded because the archive published it as probable the same day and can now say why it is not. The groom's surname is STILL not proved — the cover gives him only as «Angelo figlio di Nicola» — so that half stays open.

Caught by the 1814 processetti. · Where it lives now →

was «Two CIOFFI of strada Camellara walked to the town hall to report it.»

now LUCA COSSI, 33, a SARTORE, and GIOVANNI COSSI, 43, a contadino — both of strada Camellara. Cossi, not Cioffi.

how Re-reading the act at native resolution. Cossi is short with a double-s; Cioffi is longer and carries the -ffi ligature. At the resolution the plate was cropped to, they are not distinguishable.

Both are common Arienzo surnames and both appear all over this archive, so the confusion will recur. Recorded with the tell: at plate resolution you cannot separate them — go to the image.

Caught by re-reading at native resolution. · Where it lives now → · the plate

was Matteo Falco, head of the line: dead by 8 December 1814, with 1786 to 1814 open.

now Dead BEFORE 1809 — proved, not probable. Which puts his death in the parish register and nowhere else.

how The Arienzo death indexes for 1809-1814, read in a fraction of the time the registers would take. No Matteo Falco of the founder's generation is in them.

Published because a negative that moves a bracket is worth as much as a positive, and because it redirects the search. All six index volumes for 1809-1814 have now been read PAGE FOR PAGE — the first pass left three years partial and the claim was published as «probable» until they were finished. Taken with the marriage indexes, which also begin in 1809, BOTH of Matteo Falco's vital events lie outside the civil series: his marriage of about 1760-66 and his death are in the parish books of Sant'Andrea or nowhere.

Caught by the death indexes. · Where it lives now →

was Nothing published — this is a trap set for the future rather than an error corrected.

now A MATTEO FALCO died at Arienzo on 28 August 1809, and he is NOT the head of this line.

how The 1809 death index names his parents «Giuseppe, ed Antonia di [Guida]» WITHOUT «fu», so both were living; the same couple bury a Raffaele in 1811 and another, aged two, in 1815. The founder was born about 1740 and had children baptised from 1767.

Anyone searching «Matteo Falco death Arienzo» will land on this record, and it is exactly the shape of a find that looks like the answer. The act itself states an age and would settle it outright: Arienzo Morti 1809, an_ua14311.

Caught by the 1809 death index. · Where it lives now →

was «MARIA CARFORA», a seventy-year-old of Arienzo, 1838 atto 77

now MARIA CARBONE, widow of Saverio Carfora

how Re-read at 3400 px. She is a Carbone; the Carfora was her husband.

Maria Carfora is the single name this archive most wants to find — the probable mother of Filomena Annecchino. Unchecked it would have put a Maria Carfora born about 1768 into the Arienzo register, a century and a province away from Forchia, for someone to join to generation seven one day.

Caught by a zoom, prompted by the reading being the one I hoped for. · Where it lives now → · the plate

was «MARIA CARFORA», an eighty-year-old of Arienzo, 1837 atto 27

now MARTA CARFORA, wife of Domenico Cioffi

how Re-read at 3400 px.

THE SAME MISTAKE, IN TWO VOLUMES, ONE DAY APART. Twice the eye supplied the name it was hunting. That is no longer an anecdote about handwriting — it is a measured property of the person reading, and it is why the archive now zooms every reading that is the one it hoped for.

Caught by the same check, applied because the first failure had been written down. · Where it lives now → · the plate

was ANTONIO BERNARDO offered as «the useful counter-case» for the register's ages — 70, 76, 70, 70, three of four agreeing

now WITHDRAWN. The 1837 volume makes him FIFTY-THREE in June, eight months before the first of those seventies, on a different street.

how Read at 3400 px, 1837 atto 52.

An argument published yesterday was refuted by today's volume. Either there are two Antonio Bernardo or one man aged seventeen years in eight months. The 1838 note is corrected in place rather than deleted, because the argument was published.

Caught by the 1837 sweep. · Where it lives now → · the plate

was Generation three's wife published as ANDREANA CRISCI

now UNDECIDED. The 1839 act naming their daughter Filomena reads ANDRIANA CIOFFI, verified twice at 3000 and 3400 px.

how Both readings now stand. Either two men called Vincenzo Falco, or one of the two acts is misread.

This archive has already been caught by exactly this pair of shapes — five years of «Cossi» turned out to be «Cioffi». Settling it needs Raffaele Falco's act re-read and, if it can be found, the marriage.

Caught by the 1839 sweep. · Where it lives now → · the plate

was GIUSEPPA FERRARO filed in the register under MARIA, her middle name

now Filed under FERRARO

how FERRARO is both the Zampiello maternal surname and the Italian for a blacksmith. The builder stripped trailing trade words blind and ate the surname. The registers settle it: in 555 of 568 rows a trade follows a comma and a surname does not.

Maria Rosa Ferraro went to ROSA and Antonio Ferraro to ANTONIO. The surname filter was offering forenames. Junk surname buckets: 24 to 1.

Caught by building the Ferrara source links. · Where it lives now →

was GENERATION NINE — living, born 1955 — published with her birth date on her person page; generation eight likewise on the direct-line page

now Both masked. Living people carry a name and a relationship and nothing else.

how The person-graph builder decided redaction using a predicate that excluded ANCESTORS. That belongs in the living-relatives roster, where an ancestor is rightly not a relative — but ancestry has nothing to do with whether someone is alive, and the wife's own mother is both.

It is the exact leak the direct-line page had been fixed for once already. Verified after: no living record carries a date, place or event; no living person's node shows a year.

Caught by checking whether Carmine Falco could be published at all. · Where it lives now →

was «RAFFAELE FALCO, thirteen», dead 1838 atto 71

now PASQUALE FALCO, thirteen, son of Francesco Falco and Gasparda Falco

how Re-read at 3400 px.

Generation four IS a Raffaele Falco born in 1818. A misread Raffaele dying in 1838 would have looked like a catastrophe for the line. It is a different child of a different household entirely.

Caught by a zoom on a Falco act, as the rule requires. · Where it lives now → · the plate

was ANGELO MORGILLO, «ninety-two», widower of «Marta Grieco», 1838 atto 24

now ANGELO MORGILLO, NINETY, widower of MARIA RIVETTI

how Re-read at 3400 px on the same plate.

The wife's surname matters: Morgillo and Rivetti are the two surnames that flank this line, and here they are married to each other a generation before the founding couple's children.

Caught by a zoom. · Where it lives now → · the plate

was ANGELO DIGLIO aged SIXTY at 1837 atto 11 and SEVENTY four days later at atto 12

now SEVENTY in both — the ten-year jump was my eye, not the clerk's

how Both acts re-read at 3400 px.

It would have been the single most spectacular age in the archive. It was a misreading, and saying so is worth more than the anecdote would have been.

Caught by a zoom on a claim that was too good. · Where it lives now → · the plate

was «The register is internally consistent inside a volume and only falls apart between volumes»

now Nearly right, and now stated properly. Across six appearances in 1837 the hospital pair read: Diglio 70 six times out of six; Migliore 58 five times and 54 once. Eleven of twelve agree.

how Counting every appearance instead of two.

The first version was an overstatement built on two acts. The corrected version is the sharpest thing the sweeps can say about the ages — and it took four volumes read end to end to be able to say it.

Caught by finishing the volume. · Where it lives now →

was 1838 atto 15 dated «il dì VENTINOVE del mese di Febbrajo»

now 24 February. 1838 was not a leap year.

how THE ARITHMETIC caught it, not a second opinion. At 3200 px the word is not «ventinove»; the neighbouring acts bracket it.

A date that cannot exist is the cheapest error-detector this archive has, and it only works if the failures are written down.

Caught by the calendar. · Where it lives now →

was GIUSEPPE DI GUIDA, «ninety-five», 1837 atto 33

now NINETY-ONE — and his widow's surname LEFT BLANK on purpose

how At 3400 px the surname will not resolve between Diretti, di Ruggio and Rivetti.

Two of those are names this archive is actively hunting. Having twice today written down the name it hoped for, it recorded a blank instead. Recording a blank is cheaper than recording a wish.

Caught by the lesson of the same morning. · Where it lives now → · the plate

was «DOMENICA FALCO», wife of Raffaele Cioffi, published from the 1843 register

now FRANCESCA FALCO

how Corrected by TWO independent readings — the 1840 and 1842 volumes — not one.

The 1843 file had published «Domenica» as an explicitly flagged hypothesis. Three acts, three clerks, one spelling of her name.

Caught by the 1840 and 1842 sweeps. · Where it lives now → · the plate

was «Suor ELENA GIULIA MIGLIORE», read as a nun's name in the 1843 register

now GIULIA MIGLIORE — the word is PLENA, part of «Ave Gratia Plena», the hospital's name

how The phrase «Ospedale di Ave Gratia Plena» recurs in 1837, 1839, 1840 and 1842, which is what exposed it.

A hospital was misread as a person. It is the clearest case in the archive of a phrase being mistaken for a name.

Caught by reading four more volumes. · Where it lives now →

was Five years of «COSSI» in this archive

now CIOFFI

how The 1825 register shows the two shapes side by side in one hand.

The founding misreading of this archive, and the reason every ambiguous surname since has been left visibly undecided. One Crisci household is STILL unsettled for the same reason — «Marcello Crisci» in 1823 and «Marcello Cossi» in 1828, with the identical wife.

Caught by reading the volume act by act. · Where it lives now → · the plate

was AGOSTINO FALCONE at the head of the line

now MATTEO FALCO — and Falcone discarded

how The registers name Matteo Falco as the father in three separate children's death acts.

The family tree STILL carries Agostino Falcone as generation two's father. He is drawn on Pasquale Falco's page struck through and labelled superseded, beside the man who replaced him. Corrections here are not deleted.

Caught by the parish and civil registers. · Where it lives now →