# Aaronic ancestry and the two-worlds reading

9 September 2026. Local proposal for Mike's review. This synthesis supplements, rather than replaces, the original research records.

## The conclusion within the reading

Accepting the biblical paternal relationships, Aaron's family connects to Levi, Noah and Adam. Scripture supplies a substantial ancient family and carries priestly survival into the first century. Genetics supplies evidence that several old paternal families preserve priestly traditions across separated communities. Together, these make the continuation of Aaron's family a reasonable historical explanation to investigate. We need not first estimate millions of modern descendants.

Among the families examined, **J1-ZS222 is the best-developed positive continuity account**. Its case rests on inherited priestly tradition across communities and across several internal divisions, as well as its numerical prominence. This is a qualitative assessment, not a calculated probability. J2 and E remain serious alternative histories; an unidentified surviving family remains possible. Nothing here establishes that every modern priest is descended paternally from Aaron.

This advances the investigation of a biblical family. It does not distinguish the two-worlds reading from other accounts that preserve the same later genealogy. It neither requires nor independently demonstrates a crossing between worlds.

## Scripture and historical continuity

[1 Chronicles 24:1-6](https://www.biblegateway.com/passage/?search=1+Chronicles+24%3A1-6&version=NIV) records Eleazar's and Ithamar's houses. Sixteen and eight family heads are not a permanent population ratio. [Ezra 8:1-2](https://www.biblegateway.com/passage/?search=Ezra+8%3A1-2&version=NIV) names both houses after the exile. Thus a recent common ancestor of one modern cluster may represent a later subfamily, not Aaron himself or all his surviving descendants.

[Ezra 2:61-63](https://www.biblegateway.com/passage/?search=Ezra+2%3A61-63&version=NIV) supplies positive evidence that descent was checked. Missing credentials do not establish missing biological ancestry. The accepted count of 4,289 priests and 74 other Levites provides 4,363 ancient males conditionally on the stated paternal identities. It is not a modern lower bound.

Luke 1 and Acts 6:7 carry the accepted priestly history into the first century. [Josephus's Life, opening section](https://ccel.org/ccel/josephus/complete/complete.i.html), adds a historical claim of public paternal records and three sons born under Vespasian, after the Temple's destruction. His maternal Hasmonean connection must not become a paternal pedigree. There is no demonstrated connection here from his sons to a modern tested man.

Biological descent, eligibility for priestly service, actual office and remembered status can diverge. Nehemiah 13:28's expelled priestly descendant illustrates this distinction. Marriage does not replace a father's Y chromosome. Conversely, traditions can be acquired without the ancestry they claim. The available sources do not measure how frequently either process occurred.

## The pedigree investigation

Cherson's [2022 study](https://www.researchgate.net/publication/362906798_Historicity_of_the_Tribe_of_Levi_A_Genetic_Perspective), note 24, links the Kahana lead to a reported 74-generation ancestry report. The link resolves to [Rabbi Moshe HaKohen's Geni profile](https://www.geni.com/people/Rabbi-Moshe-HaKohen/6000000169041861168). Its accessible response exposes an access/security page, not the chain. The related [Geni discussion](https://www.geni.com/discussions/269016?msg=1826033) was also inaccessible in this inquiry.

A [public lead catalogue](https://judaism.stackexchange.com/questions/23190/can-any-kohanim-trace-all-the-way-back-to-aaron) describes a 98-generation Kahana compilation and says its research is not fully verified. This is a secondary description, not direct access to Kahana's sources. Different generation counts may concern different endpoints or compilations; they are not by themselves a contradiction.

**The exact gap:** the generation-by-generation chain, documents supporting its links, and a documented connection between the pedigree and sequenced men. Without these, paternal continuity, female links, identities, missing generations and chronology cannot honestly be reported as audited. Agreement among recent relatives could support the recent pedigree without validating its remote extension. This is a gap in a later genealogical claim, not a demand to prove the biblical genealogy afresh.

Concrete additional leads were followed:

- The cited genealogy in *Eitan Haezrachi*, [scan page 216](https://hebrewbooks.org/pdfpager.aspx?hilite=&pgnum=216&req=589&st=), and *VeTzivah HaKohen*, [scan page 72](https://hebrewbooks.org/pdfpager.aspx?hilite=&pgnum=72&req=3727&st=), returned HTTP 403. Reports of their contents remain leads, not inspected pedigree evidence.
- [Sheilat Yaavetz I:89:5](https://www.sefaria.org/Sheilat_Yaavetz,_Volume_I.89.5) was read through Sefaria's public text API. Emden discusses the sufficiency of a court-certified family pedigree reaching Temple times, but says such a document had not reached him. This supports the historical importance of pedigrees, not a newly identified surviving one.
- Marciano's [Cohen-Tzaban family account](https://moreshet-morocco.com/2012/08/10/%D7%99%D7%97%D7%A1-%D7%93%D7%91%D7%93%D7%95-%D7%90%D7%9C%D7%99%D7%94%D7%95-%D7%A8%D7%A4%D7%90%D7%9C-%D7%9E%D7%A8%D7%A6%D7%99%D7%90%D7%A0%D7%95-6/) gives a sequence of fathers and children from an early modern Aaron Cohen-Skali. It expressly leaves some descendants unknown and includes daughters. The opening Hebrew and parenthetical civil date disagree (5302 versus 1560), so that date needs checking. It supplies no sequence of links to biblical Aaron and no tested-man identifiers. This is useful family history, but cannot currently assign an Aaronic haplogroup.

No family members or researchers were contacted. Further enumeration of claims with the same missing connections would not strengthen identification. The inaccessible documents are recorded so the investigation can resume if they become available.

## Genetic reconciliation

[Sahakyan et al. (2021)](https://www.nature.com/articles/s41598-021-85883-2), supplementary table S2, explicitly identifies **B877 = S8223 = ZS222**, at hg19 position 2709209 / hg38 position 2841168, T→C. Older node118 groups 29 mutations that newer representations separate. The alias identity therefore does not make every historical branch assignment equivalent to today's narrow branch.

Its supplementary note describes the Imperial-period individual listed as R835 in table S3: two derived and seven ancestral calls among nine sequenced positions, potentially a sister branch of the Jewish family. The table places him at Civitanova Marche, 27 BCE–300 CE. He is **not an identified ancient priest**. The note does not specify those nine positions. Without their calls, the exact modern placement remains unresolved. This is why the old B877 label does not by itself contradict the newer claim that no ancient CB-01 individual is known.

The attached 2025 Lipson preprint and supplements were previously audited in [the supplement review](../lipson-cohanim-review/SUPPLEMENT-REVIEW.md). Their ZS222 block is not identical to the sanctioned dump's block. The corrected all-occurrences tally is 17 of 25 names at J-Y3088, two at its immediate parent J-S20075, and six unmatched names. FGC13874 is explicitly present at J-Y3088 as `Y3088/FGC13874/Z18270`, as well as at a second node under J2. My earlier description wrongly omitted the J-Y3088 occurrence and understated the agreement with ZS222. The six unmatched names remain unresolved after checking slash-separated aliases; absent names do not prove absent mutations. B877 is not itself a defining SNP name in the dump: its documented alias S8223 provides the bridge to `S8223/ZS222` at J-Y3088. Do not pick whichever tree date produces a preferred biblical fit.

The local J-Y3088 interval and Lipson's interval must remain separate. Using the approximate **2025 CE reference-year subtraction**, 2025 − 2800 = −775, giving about **775 BC**; 2025 − 3200 = −1175 and 2025 − 2300 = −275 give **1175–275 BC**. The older endpoint is 91 years after the site's 1266 BC Exodus, so this interval does not include that date. Lipson Table 3 reports **834 BC, interval 1291–440 BC**, which does include it; the prose's 833 BC and 1289–440 BC also include it. The older clade-denominator frozen window implies a 2026 reference under direct subtraction instead: 2026 − 2800 = −774, with endpoints 2026 − 3200 = −1174 and 2026 − 2300 = −274. That one-year difference also leaves the Exodus outside. These are separate estimates, not one combined interval. A younger common ancestor can represent an Aaronic subfamily; overlap alone does not identify Aaron.

The proposed identification of S17446 and FT157560 as Eleazar's and Ithamar's separate houses conflicts with the newer ordinary-clock date of their shared ancestor, J-Z18271: about 687 BCE, interval 1122–312 BCE, later than the accepted Exodus of 1266 BCE. They could instead be later subdivisions of one house. This challenges the two-house assignment, not the wider J1 continuity account. Ordinary-world dating uncertainty deserves examination, but the ancient-world clock cannot explain away every much later date.

## Complete competing histories

| Account | Positive case | Outstanding historical work |
| --- | --- | --- |
| J1-ZS222 preserves part of Aaron's family | Broad priestly tradition and substantial internal diversity across communities | Explain the other recognised priestly clusters and connect this family's earlier ancestry; identify neither Aaron nor Noah with its younger sampled common ancestor |
| J2-CB03 preserves part of Aaron's family | Old dispersed family with a substantial priestly western division | Resolve eastern affiliations and explain the extensive J1 priestly tradition through another history |
| E-CB02 preserves part of Aaron's family | Priestly witnesses in both major divisions and several communities | Connect its founder to Aaron and explain the other deep priestly families |
| Another line survives, with or without recognised status | Loss of status is compatible with paternal survival; sampling is incomplete | Identify that line and explain why the conspicuous sampled priestly families represent other ancestry |

These histories are not declared equally plausible. J1 currently combines the most developed positive evidence in this comparison. Nor are they exhaustive haplogroups of Noah: other Noahic sons and later paternal families could survive outside the priests. Sample size alone does not settle which account is true. Most of the later expansion of any candidate is a shared nuisance process, not independent evidence of its founder's biblical name.

## What can be calculated

[check.py](check.py) reproduces the alias check and the previous local-tree time cuts. It reads the saved publisher supplement and the sanctioned dump, records hashes, and makes no network calls. With an alternative dump, pass `--tree PATH`. It requires Python and openpyxl. Results are in [checks.json](checks.json).

At a cut of 2,000 in the dump's years-before-present units, J-Y3088 has 37 represented paths, or 22 requiring a descendant SNP-defined terminal; E-BY8508 has 3 or 2; J-FGC4975 has 3 or 2. At cuts 2,200 and 1,800, the SNP-supported counts are respectively J1 19 and 22, E 2 and 2, J2 2 and 4. These are sensitivity checks using point dates, not confidence intervals. They establish represented old diversity conditionally, not the number of living men, their priestly identities or an unconditional minimum of Noah's descendants.

For any candidate H, posterior odds require prior odds multiplied by P(data|H)/P(data|alternative), integrating uncertainty about survival, recognition, migration, sampling and dating. The available evidence does not estimate those historical processes sufficiently to provide defensible identification probabilities. The [minimum-status-change exercise](../priestly-status-tree/REPORT.md) helps assess inherited priestly status. An Aaronic root and a non-Aaronic root already carrying the same status can have identical subsequent likelihoods in that model. It cannot manufacture an identification by renaming a root.

## Implications and stopping point

A securely connected modern Aaronic descendant would also be a Noahic and Adamic descendant on the accepted genealogy. If he lies within J1, that places their paternal ancestry on the ancestral path leading to that family, subject to the stated inheritance model. It does not identify Noah with ZS222 or locate Adam on a dated stem. Surviving descendants of the original man whose chromosome was matched, and possibly Cain's descendants, further complicate identifying a unique Noah clade from sequence alone.

Such a connection would favour compatibility with the ordinary human Y tree. It would not require an exact match to a particular chromosome present at Adam's creation. Matching remains possible. A sequence designed to fit that ancestry, ordinary copying, and creation of an identical sequence need not be distinguishable through descendants. Nor does lack of an identified outsider chromosome force matching without a defensible survival and detection model.

The accessible evidence now supports a positive qualitative continuity account and conditional diversity results. It does not support a worldwide modern lower bound or numerical Noah-candidate odds. Further useful progress requires specific new evidence: the cited pedigree with supporting records and tested-family links; ancient per-site calls at the disputed branch; or individual-level affiliations, sequences and dating inputs for the newer study. These are precise gaps, not reasons to repeat another haplogroup filter.