Method

Every claim on this site carries a confidence and a source. The point of saying so is that it lets a claim be overturned.

Agostino Falcone, and what replaced him

The line was previously carried above Pasquale Falco to an Agostino Falcone born about 1740, with a wife Elisabetta Sabina . That link has now been not merely doubted but superseded. Pasquale's own baptism has been found.

Anno Domini millesimo septingentesimo sexagesimo septimo, die vero vigesima quarta … baptizavi infantem natum ex Mattheo Falco, ed Francisca Crisci coniugibus huius archipresbyteralis collegiatae ecclesiae Sanctae Andreae Apostoli Terrae Argenti, qui nominatus est Paschalis. — Arienzo baptisms, 24 October 1767

The father is Matteo Falco , not Agostino Falcone. The reasons the Falcone claim was doubted before the baptism turned up are worth keeping, because they are the reasons it should have been doubted:

• The surname was different — Falcone, not Falco.
• Falcone is not an Arienzo surname. Across 361 register pages mentioning a Falco, Falcone appeared on two, both index lists. Crisci appeared on 113, Morgillo on 76.
• The one Falcone in a register entry is the officiating priest, not a parent.
• No record was ever produced for him or his wife.
• FamilySearch's index conflates the two names, returning both for either query — verified: identical result counts across four separate collections. A researcher can acquire a Falcone ancestor from that index without ever choosing to.

Matteo Falco is now the head of the documented line. Above him the baptism register has a gap between 1738 and 1761, and his own baptism most likely falls in it.

On dates read by machine

Much of what is new here was found by handwriting recognition applied to registers no one has indexed by hand. It is a remarkable tool and it is wrong about digits more often than about names. Where a machine-read date is load-bearing it is flagged, and it should be confirmed against the image before it is believed. One such case is on the Arienzo page.

Three wrong negatives in a single day

What replaced them

• Find out what exists before concluding it doesn't. antenati.cultura.gov.it/search-registry/?localita=<town> returns the ark of every volume a town has. Guessing arks from an arithmetic rule can only ever find what you already know about.
• Read the index first. A year's deaths are an index of one to four images listing every death with both parents named, against fifty images of register. The acts still carry what an index omits — the street, the hour, the declarants, the word «parente» on which this archive's best finds turn — so the register is not obsolete. It is just not the place to start.
• Run a new method against answers you already have, before you trust it. The first control test of the index method reproduced a known death — and caught two errors in the known answer: a date misread by ten years' worth of a single word, and two witnesses given the wrong surname.
• A “do not search again” note must carry the reasoning that produced it, so the reasoning can be checked. A wrong negative is the most expensive mistake an archive can make, because nobody ever re-tests it.
• An absence in a finding aid is not a negative. The 1834 index is a single image covering A to C. A name beyond C is not missing — it was never photographed.

Five things the Arienzo registers taught us to distrust

Reading twenty annual volumes act by act turns up habits of the source that are easy to miss from a single record, and each of them qualifies claims this archive itself has made.

A declarant's street is not where he lived

On 20 June 1823 Giuseppe Falco declares two deaths on the same day. In one he is «domiciliato… strada Camellara», in the other «domiciliato… strada Sant'Andrea». He had not moved between breakfast and supper. In this printed form the declarant described as vicino — neighbour — is written at the street of the house where the death happened. So a declarant's street records where the dead person lived, not where the declarant did. Anywhere this site places a person on a street from a declarant line alone, that is the weaker kind of evidence.

Ages move by a year a day

The same Giuseppe Falco is forty-three in one of those two acts and forty-two in the other — same clerk, same afternoon. Across nineteen months his stated age implies birth years from 1780 to 1784. Ages in these registers are an approximation offered by a neighbour, and this archive treats a single stated age as a range, never a date.

Forenames repeat until they are useless alone

Arienzo held at least three Carlantonio Morgillo, three Maria Falco, two Paolo Rivetti, two Marcello Rivetti, two Nicola Morgillo, two Francesco Quotolo and two Arcangelo Crisci — all alive or dying within a decade of one another. Identity here rests on the combination of parents, spouse, street and trade. A matching forename is where an identification starts, never where it ends.

Two surnames can be one letter apart, and the clerk knew the difference

From 1823 this archive recorded a household as «Francesco Cossi and Maria Falco» of strada Camellara — the house whose five dead children Giuseppe Falco declared, and the strongest Falco kinship lead outside the direct line. It is CIOFFI. Read at full resolution, the 1825 volume writes the Camellara family with a dotted i after the C — Cesare Cioffi, Francesco Cioffi, Giuseppa Cioffi — and, in the same hand and the same book, writes the unrelated family of strada Cossi without one: Domenico Cossi, Crescenzo Cossi. The clerk distinguished two families; this archive conflated them for five years across six annual registers.

The correction is published rather than swapped. The reading logs keep the original text and carry a banner saying what was wrong; the plate showing the four words side by side is on the Documents page. Two lessons follow. The first is procedural: a surname read at strip resolution is a hypothesis, and only a full-resolution reading of a distinguishing pair — the same clerk writing both candidates — settles it. The second is that the same doubt is still open elsewhere — or was. The priest buried in 1827 as Don Giuseppe Cr-i-sci or C-o-ssi, son of Andrea, possidente, and Orsola Zimbardo was settled sixteen years later by an act nobody was looking for: in June 1843 the register buries Celia Cioffi , eighty, «figlia di ANDREA CIOFFI defunto, di professione POSSIDENTE … e di ORSOLA ZIMBARDO defunta». Same father, same trade, same mother, spelled plainly. He was Don Giuseppe Cioffi. Which is the third lesson: a hand that will not resolve is sometimes resolved by a different clerk in a different decade, and the way to catch that is to read whole volumes rather than hunt single acts. One reading is still open — the household held as Marcello Crisci & Maria Rosa Mauro, which 1828 writes as Marcello Cossi with the identical wife.

Five: a name is not a person

Reading 1840, 1841, 1842 and 1843 end to end — three hundred and thirty acts — turned up something the archive had been half-aware of and had never had to face at scale. The same name belongs to several people at once, on one street, in one year. In the 1840 volume alone:

• Three men called Sebastiano Lettieri on strada Costa — with wives Felicia Cioffi in August, Margarita Porrino in November, and Mariangela Domino at a death in 1841. Three women in fifteen months is not one man remarrying.
• Two Gennaro Lettieri, two Clemente Zimbardo a week apart, and two Pasquale Morgillo a generation apart — each pair separated only by a wife's name.
• A third Gennaro Morgillo and a second Marco Morgillo, both known only as dead husbands in other people's acts.

The ages do not rescue you. Giuseppe Falco is fifty-six on 16 October 1841 and fifty-three on the 19th, in the same hand. Angelo Diglio runs 72 · 73 · 73 · 77 · 72 across eleven months. Luca Cioffi is thirty-three in July 1841, thirty in September, thirty-three again in 1843 and thirty-four the following day, when the clerk also changes his trade. Two acts in the 1840 volume give one couple two daughters dying eleven and twenty-two days old, seven weeks apart — births eight weeks apart, which cannot happen. Both are published as read, with the contradiction stated.

What follows for anyone using this archive: a matching surname is not a match, and a matching surname with a plausible age is still not a match. What joins two records here is a named parent, a named spouse, a stated relationship, or a street plus a trade repeated across volumes — and where the archive has only a name, it says so and leaves the person unjoined. The Morgillo table is published for exactly this reason: eleven Morgillo who do not connect to this family, set out to show how little a Morgillo in a record proves.

What is not known

An archive that only publishes its findings is half a record. Everything this one has failed to establish — and every search that came back empty — is set out at Open questions, including the ground already covered, so that the next attempt does not repeat it.

What a DNA test can and cannot reach

Autosomal DNA halves at every generation, and the reporting floor on a consumer estimate is about one per cent. Six generations back an ancestor contributes about 1.5% of a descendant; seven back, under 1% — beneath what the test reports at all.

For this family the arithmetic lands somewhere pointed. Generation eight was tested, so generation two sits at 1.56% and generation one — Matteo Falco, the founding ancestor of this entire archive — sits at 0.78%, below the floor. The paper reaches nine generations; the DNA reaches about six.

They are therefore not two views of the same distance, and the consequence is a rule: an absence in a DNA result is a fact about the test, not about the family. Where the two overlap they agree exactly. Beyond the overlap the test is silent, and silence is not evidence. The full working is on the DNA page.

How a person page is built, and what it refuses to do

A person page used to be built from the register households alone. Those are excellent on Arienzo before 1850 and nearly empty after it, because the later generations lived at Arpaia and Brisbane. So Angelo Zampiello's page showed one box — him, in his father's house, with no wife and no children — while the Ferrara page three clicks away named his wife and five of their children. That was fixed on 10 September 2026.

The pages now merge the registers with the family tree. The tree is unverified, and the answer was not to choose between them but to merge and say which is which on every single edge:

An edge keeps every source that carries it, so a reader can see when two independent sources agree — Pasquale Falco's marriage to Chiara Rivetti is marked line + register + tree, and that is worth more than any one of them.

What the merge refuses to do

Joining records is where an archive invents people, so the joins are narrow and each one is named on the page it affects. Every rule below exists because it caught a real error while this was being built:

• A witness is never a relative. A declarant is the man who walked to the town hall, very often a neighbour. The whole value of the 1828 «parente della defunta» find is that the register says so when he is kin. Turning every witness into family would throw that away.
• A daughter is not a parent. The household files use child, but also daughter and son; reading only the first put real children in the parents' row, which is how Chiara Rivetti briefly had her own father and mother listed as her husbands.
• One woman, not two. A woman is a daughter in her father's house and a wife in her husband's, so she arrived as two records and her husband's page listed her twice, as two different wives. The two are joined only when the name leaves no room — exactly one child of that name in the whole file and exactly one head — and then only if the join passes the next test.
• Nobody may become their own ancestor. Vincenzo Falco and Andreana Crisci called a son Pasquale. Joining that child to the one head called Pasquale Falco made generation two his own grandson — a dead infant promoted into the direct line by a forename. The name is only ever a candidate; the test is that the parent graph must stay acyclic, and a parent must be an earlier generation than their child.
• A generation is a couple, not a forename. Matching generations by name alone labelled four different Matteo Falco "Generation 1", including the possidente of strada Porta di Sopra whom this archive has always called unjoined. A generation is now identified only when the ancestor and their spouse both agree.
• Where the name repeats, nothing is joined at all. The tree holds four Angelo Falco, four Francesco, three Giovanni and three Crescenzo. None of them is merged with anything. They are drawn separately and the reader is told why.

The builder refuses to write the file at all if any of this fails: no cycles, no marriage only one party is in, no parent without the child back, and no date on a living person.

The claim this archive already withdrew, still on the page

The family tree gives Pasquale Falco of generation two a father called Agostino Falcone — the ancestor examined and discarded at the top of this page. Drawing him without a word would quietly put a retracted claim back on the site. He is drawn, struck through, and labelled superseded, beside the Matteo Falco who replaced him. Corrections stay visible here; they are not deleted.

And the tree's own spine stops one generation short

The nine generations were checked by walking the family tree upward from the living generation and comparing each step with what this archive publishes. They agree for eight generations and then part company: at the ninth step the tree still reaches Agostino Falcone. So generation one is taken from the registers alone, and the tree's own head of line is left unlabelled. The disagreement is the check working, not a fault in it.

What the links mean

A link on this site never claims more than the archive knows. Names in the prose became clickable on 10 September 2026, after the question was asked why a person named on a family page did not take you to that person. The answer had been simply that nobody had built it. Building it raised a problem worth stating, because the obvious version would have been dishonest.

This parish repeats its forenames relentlessly. There are three Maria Falco, four Pasquale, two Raffaele Falco inside the direct line itself — generations four and six — and two Carmine Antonio, generations five and seven. Sending a reader from the words «Raffaele Falco» to one particular Raffaele Falco would be an identification the archive has not made, dressed up as navigation. So the rule is:

Sources are linked the same way. Where a claim rests on a record, the citation itself is the link — to the plate on Documents, or out to the Antenati or FamilySearch entry. Where a claim rests on something that cannot be linked — a paper estratto in the family's own scans, a cemetery search that returns no stable URL, an unverified entry in the family tree — the citation says so in place of a link, rather than leaving the reader to assume there is a record behind it.

Living people

This rule changed on 10 September 2026, at the request of the family whose archive this is. It is recorded as a change rather than presented as though it had always been so.

Until then, living people were omitted from the build entirely — not hidden, not gated, not present in the output. They are now named, because a record of a family that stops at the dead is not a record of a family, and the living are the reason anyone reads it.

What is published for a living person is a NAME, and their relationship to the family. Nothing else. No date of birth. No date of death — they have none. No birthplace, no residence, no age, no occupation, no photograph, no record link. Those fields exist in the family tree and are dropped on the way in, by the tool that builds the list, not by hand.

Three filters decide who counts as living, because the family tree's own flag cannot be trusted: it means «no death date recorded», which had it marking a man born in 1832 as alive. Entries named Unknown … are placeholders for a gap, not people. A birth before 1920 means dead whatever the flag says. And anyone in the direct ancestral chain is by definition not living. Of 137 flagged, 104 are published and 33 were filtered out.

A removal request is honoured within days, without argument and without requiring a reason — and that promise now matters more than it did.