# Research Architecture for the Jacob Historical Model

Version 0.1 · 29 September 2026 · Research architecture and first methods diagnostic

## Purpose and current status

This project asks how widely the genealogical descendants of one stipulated ancient Levantine individual could spread under historically plausible conditions. Jacob is the simulation’s specified founder. His existence, family size, chronology and descent of any modern group are not historical conclusions of this model. Genealogical descent is inherited through parent links; identifiable DNA is not required.

The architecture below is complete as an initial research specification. Historical calibration is not complete. It defines 28 geographic containers, 32 priority edge families, 16 research regimes, a parameter contract, three alternative interpretations and a source ledger. Quantitative ancient rates and actual activation dates remain null where they have not been established. The edge list is a prioritized research skeleton, not an exhaustive world network. Missing candidate edges cannot be treated as demonstrated historical isolation.

A separate, small stochastic pedigree diagnostic accompanies this architecture. Its inputs are entirely grade D experimental choices. It tests mechanics and sensitivity in synthetic populations; it cannot produce an AD 2026 forecast, a historical confidence interval, or a claim of universal descent. The complete Historical Adaptive — Monte Carlo historical engine remains subject to the staged acceptance criteria below. Keeping that distinction visible is part of the model.

## 1. Retain the Rohde framework and change the historical inputs

Rohde, Olson and Chang’s computational model nests local mating towns within intermediate regions and larger geographic regions. It has overlapping lifetimes, mostly local mate choice, distance-sensitive migration, dated interregional ports and concentrated immigrant settlement. Supplement B explicitly treats port magnitudes as rough estimates. Its adult-survivor representation and restricted simulated population also require care when transferring results to real headcounts. [Rohde et al. (2004)](https://doi.org/10.1038/nature02842), [Supplement B](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementB.pdf).

The paper’s simpler mathematical graph model is a separate benchmark. Reproducing graph-model MRCA times does not validate an individual historical simulator or establish the probability that a particular founder survives. Retain the existing Rohde-style sensitivity presets as comparators; make Historical Adaptive a separately versioned model. [Supplement A](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementA.pdf).

Proposed implementation:

- Keep individual parent links, birth/death dates, town, geographic residence, social participation and a Boolean founder-descendant tag.
- Set `child.descended = mother.descended OR father.descended`. Never sum pedigree paths into people, infer the tag from ethnicity, or let the tag change fertility or migration.
- Treat geographic macroregions as accounting containers. Local towns and social populations are the mating units.
- Preserve short-distance movement and inverse-distance kernels; allow rarer long-distance routes, changing transport, opening bursts and subsequent lower flow. Calibrate their quantities by event or generation, not invented annual measurements.
- Separate departure, arrival, settlement, reproductive opportunity, cross-group parenthood and host participation. An individual birth engine realizes reproduction through age/partner/birth events; do not additionally multiply it by a second “reproductive success” factor measuring the same thing.
- Model household movement and correlated mortality, especially for deportation, slavery, war and colonization. Independent person-by-person transfers can exaggerate the number of independent bridges.
- Use annual internal event scheduling if useful, while retaining the actual coarse evidence intervals. A fine computational time step is not fine historical knowledge.

## 2. Define the founder experiment before historical calibration

Initialize one specified individual around 1800 BCE in a southern-Levantine local population. A birth date, age, sex, initial household and local social setting must be explicit model assumptions. Do not assume a Jewish population exists under that label in 1800 BCE, make all later Israelites descendants, or guarantee twelve surviving sons. An ordinary founder, a founder conditioned on parenthood and a pre-established family are distinct experiments.

The single-founder default should be unconditioned on future reproductive survival. A separate conditional view may answer what happens if the lineage survives. Simulate local diffusion before sampling later emigrants/deportees. If only a small part of the source is descended at the time of an event, a historically attested migration may carry no tagged person at all.

Use a demographic initialization/burn-in. If kinship restrictions are modeled, generate enough prior pedigree to make those restrictions meaningful. Do not pretend that this synthetic initialization reconstructs actual Bronze Age genealogy.

Report:

1. Probability of any living descendants at AD 2026, with extinct runs included.
2. Regional and global descendant-count/share distributions, both unconditional and explicitly conditional on survival.
3. Separate times of first arrival, first locally born descendant, first cross-group descendant child, and first sustained host branch.
4. Town/community coverage and residual unreached populations, alongside population-weighted averages.
5. Unreached destinations as right-censored by 2026. Never average only reached cases and call that the average arrival time for all runs.

A universal-descent claim requires checking every represented living person and meaningful residual barriers. A sampled town or continent average cannot establish descent of every actual human. “Some person was a universal ancestor” and “this specified person was a universal ancestor” are different questions.

## 3. Geographic design and resolution

Start with the 28 containers below. Each includes nested regional towns and social networks, not one randomly mating population. These partitions are grade D model choices to be tested, not historical claims that the named regions were closed populations.

| ID | Geographic container | Required subdivision or trigger |
|---|---|---|
| LEV_S | Southern Levant | Highest resolution from initialization: hill country, coastal/lowland networks, Transjordan and relevant community distinctions; never treat a political kingdom as a closed pedigree. |
| LEV_N | Northern Levant and Cilicia | Split coastal ports, inland Syrian cities and upper Euphrates corridors when deportation and diaspora edges are introduced. |
| EGY | Egypt and Nile corridor | Alexandria/Delta, Nile valley and Nubian connections; separate urban diaspora from host settlements. |
| MES | Mesopotamia | Upper/lower Mesopotamia and Jewish diaspora/host strata. Capitals and deportee settlements can be separate port localities. |
| IRC | Iranian plateau and Caucasus | Retain distinct nested Iranian and Caucasian corridors from first use; mountain isolates are not mixed with lowland crossroads. Promote to separate top-level regions if route sensitivity requires. |
| AAB | Anatolia, Aegean and Balkans | Ports and imperial capitals separate from rural hinterlands; Aegean and continental routes can have different opening histories. |
| ARB | Arabia | Northern caravan corridor, western/Red Sea settlements and southern/Gulf ports; separate nomadic, oasis and maritime strata as needed. |
| ITM | Italy and central Mediterranean islands | Rome and major ports versus hinterland; later overseas emigrant source populations. |
| MAG | Maghreb | Coastal cities, interior and Saharan gateways; no automatic mixing across the desert. |
| IBE | Iberia | Religious/community boundaries, conversion and diaspora branches; later Atlantic ports. |
| WNE | Western and northern Europe | Initially nested regional towns; refine British Isles, Atlantic seaboard and inland source regions before transoceanic migration. |
| ECE | Eastern and central Europe | Jewish settlement zones, host strata, Baltic/Black Sea routes; refine when founder or expulsion history matters. |
| CAS | Central Asian oases, steppe and Siberia | Distinct corridor and sparsely connected hinterland classes; split Bering/Arctic before considering an American connection. |
| SAS | South Asia | Never a panmictic node: regional town networks plus endogamous social strata. Refine northwest land routes and west/east coastal receiving populations first. |
| EAM | East Asian mainland | North/south and interior/coast nested regions; refine Central Asian corridor destinations before giving them access to all mainland towns. |
| KJI | Korean peninsula and Japanese archipelago | Separate regional populations under a shared accounting container; explicitly gated maritime crossings. |
| MSEA | Mainland Southeast Asia | River/lowland towns and upland populations separated; connect Indian Ocean and East Asian ports selectively. |
| ISEA | Island Southeast Asia | Island chain structure; no single regional mate pool. Split western maritime ports from eastern islands before outward routes activate. |
| HEA | Horn of Africa and East African coast | Separate Horn/highland, Swahili coast and inland networks; coastal mixture does not imply continent-wide entry. |
| SWA | Sahel and West Africa | Saharan corridors, inland populations and Atlantic source coasts; refine the latter before forced Atlantic migration. |
| CSA | Central and southern Africa | Keep subregional local networks; split southern colonial ports, central source zones and less connected groups when relevant. |
| MAD | Madagascar and western Indian Ocean islands | Separate settlement sources and later local substructure; island-specific founding histories. |
| NAM | North America and Arctic | Initially disconnected/coarse internal networks; split Arctic, Atlantic contact zones and interior populations before relevant contact. |
| MCA | Mesoamerica and Caribbean | Island/continental structure; split colonial receiving ports, Indigenous groups and plantation/settler populations after contact. |
| SAM | South America | Coastal Atlantic/Pacific and interior networks; preserve sparsely connected Indigenous populations. |
| AUS | Australia and Tasmania | Distinct Indigenous regional networks and, later, immigrant/settler populations; contact with one coast never covers the continent. |
| PNG | New Guinea and Near Oceania | Highland/lowland/coastal and island structure maintained even at low research resolution. |
| OCE | Remote Oceania, including Aotearoa/New Zealand | Island graph and settlement timing; split islands with different contact dates before they are targets of transmission. |

**Highest resolution, 1800 BCE–AD 700:** southern-Levant highlands/Judah–Jerusalem, Samaria, Galilee–Jezreel, coastal settlements and Transjordan; northern Levant urban/coastal corridors; upper versus lower Mesopotamia; Alexandria/Delta versus the Nile/Elephantine context. Split Cyrenaica from the wider Maghreb before applying the 115–117 shock. Track enclaves separately inside Egypt, Anatolia/Aegean and Italy. Distinguish northwestern Arabia from Himyar, and Iranian lowland gateways from Caucasian mountain networks.

**Medium resolution after entry:** Mediterranean hosts and hinterlands, European receiving populations, Mesopotamian/Iranian hosts, Arabian corridors, Caucasus/Central Asia, South Asian coastal gateways and their internal social boundaries. A host network earns research priority when descendants can reach it, not just because it has a large population.

**Initially coarse:** distant American interiors, Australian regional networks, New Guinea, many Pacific islands and other weakly connected regions may have sparse parameter histories before plausible contact. Coarse must still preserve islands, uplands, coast/interior and social barriers. If the outcome is universality, even tiny isolates require representation from the outset. They cannot safely be merged into continental averages.

Split a region when a distinct port, social barrier, different route timing or sensitivity result changes the answer. Conserve people, age structure and tagged individuals when splitting. Compare finer and coarser runs against Monte Carlo noise and a predeclared practical tolerance (for example, coverage or arrival-delay tolerances chosen for the research question). Merge only when the distinction has negligible effects across relevant scenarios, including outcomes for small populations.

## 4. Historical regimes and local change points

The following are research windows. They do not impose simultaneous global changes. Within a window, activate or modify individual routes only when source extraction supports that choice; store uncertain opening intervals and attestation dates separately.

| Window | Research regime | Genealogical reason |
|---|---|---|
| 1800–1200 BCE | Founder and Bronze Age background | Local survival and early boundary permeability; ethnic labels cannot be projected backwards. |
| 1200–746 BCE | Early Iron Age reorganization | Test settlement-based local structure; no guaranteed Israelite founder tag. |
| 746–597 BCE | Assyrian movement | Multiple campaigns, incoming settlers, forced family migration and local continuity. |
| 597–539 BCE | Babylonian displacement | Separate 597 and 587/586 BCE event windows, Judah survival and eastern settlement. |
| 539–332 BCE | Persian mobility | Return routes coexist with families remaining in Babylonia; Egypt differs by locality. |
| 332–63 BCE | Hellenistic diaspora | Egyptian, Anatolian and Aegean communities; incorporation/conversion alternatives. |
| 63 BCE–AD 66 | Early Roman networks | Pre-70 diaspora, Rome, urban hosts, captivity, manumission and ordinary migration. |
| AD 66–135 | Three distinct crisis windows | 66–70, 115–117 and 132–135; localized loss and relocation, not one total expulsion. |
| AD 136–699 | Late-antique boundaries | Regional Christianization, Sasanian contexts, Galilee, Arabia and changing institutions. |
| AD 700–1199 | Imperial and maritime corridors | Host mobility, Indian Ocean, European enclaves and conversion; merchant residence is not mixing. |
| AD 1200–1491 | Corridors, shocks and island contacts | Mongol-era routes, plague, European expulsions, pre-European Pacific–American alternatives. |
| AD 1492–1799 | Atlantic and oceanic reconfiguration | Directed forced migration, colonization, conversion, Indigenous barriers and local collapse. |
| AD 1800–1913 | Mass migration and urbanization | Destination communities, rail/steam, indenture and host-entry delays. |
| AD 1914–1945 | War and catastrophic displacement | Family-correlated losses, genocide, forced relocation and closed borders. |
| AD 1946–1989 | Postwar migrations | Israel, refugees, decolonization and changing host marriage markets. |
| AD 1990–2026 | Modern migration and boundaries | Separate travel, residence, fertility and intermarriage; limited generations remain. |

Keep the Assyrian campaigns as multiple windows rather than one “722” switch. Preserve the 587/586 BCE ambiguity where relevant. Model 66–70, 115–117 and 132–135 CE as distinct, geographically uneven shocks. Jewish diaspora networks already existed before 70; no wholesale post-70 expulsion is a default. Late-antique religious/legal developments are regional hypotheses, not universal mating-rate switches. [Radner](https://epub.ub.uni-muenchen.de/58863/), [Smallwood](https://www.cambridge.org/core/books/abs/cambridge-history-of-judaism/diaspora-in-the-roman-period-before-ce-70/AD1E5333322627A5BD5063DB966A0813), [Raviv & Ben David](https://www.cambridge.org/core/journals/journal-of-roman-archaeology/article/cassius-dios-figures-for-the-demographic-consequences-of-the-bar-kokhba-war-exaggeration-or-reliable-account/2D9838573C499EFC816F5E1569E63B58).

Later important research windows include Islamic imperial networks, medieval Indian Ocean household connections, European settlement/expulsion/conversion, Mongol-era relocations, epidemics, Atlantic colonization and slavery, industrial migration, genocide and postwar resettlement. Political unity and shipping intensity are covariates to investigate, not automatic reproductive bridges.

Recent evidence supports localized pre-European Pacific–American reproductive contact. Include an alternative topology; do not insist that 1492 is the first possible American connection. Also do not assume that an ancient Levantine tag had reached the relevant Pacific source before such a crossing. Direction, locality and time-respecting paths all matter. [Ioannidis et al. (2020)](https://doi.org/10.1038/s41586-020-2487-2), [Moreno-Mayar et al. (2024)](https://www.nature.com/articles/s41586-024-07881-4).

## 5. Priority connectivity graph and evidence ledger

Every row below is an edge family to research. Bidirectional candidates must become two independently parameterized directed records; a reciprocal line on the research map is not equal flow. Same-region rows represent social boundaries. A/B/C grades below concern the broad event or contextual evidence. All unestimated numerical rates remain D; a strongly attested movement does not make its intermarriage probability strongly measured. D route grades identify design/benchmark candidates still needing independent historical support.

| ID · route | Historical evidence | Model inference | Unresolved choice · sources |
|---|---|---|---|
| E01 · Local Levantine household boundaries (LEV_S → LEV_S; evidence B) | Sampled Bronze/Iron Age movement and admixture; later household evidence varies. | An early outward reproductive link may unlock larger networks. | Founder survival, town size and rates unmeasured; do not project later Jewish law backwards. [J01](https://pubmed.ncbi.nlm.nih.gov/32470400/), [J06](https://www.jstor.org/stable/j.ctv346sfv) |
| E02 · Assyrian resettlement (LEV_S → MES; evidence B) | Eighth-century imperial movements and some destinations are documented. | Forced family movement can carry a surviving tag east. | Deportees must be sampled from actual simulated source prevalence. [J02](https://epub.ub.uni-muenchen.de/58863/) |
| E03 · Incoming settlers in Samaria (MES → LEV_S; evidence B) | Resettlement also brought outsiders into conquered territories. | Incoming untagged families may connect source and host pedigrees later. | Continuity, selection, numbers and mixed unions unresolved. [J02](https://epub.ub.uni-muenchen.de/58863/) |
| E04 · Judah to Babylonia (LEV_S → MES; evidence B) | Displacement and later Judean family residence are attested. | Simultaneous source bottleneck and eastern reservoir. | Separate 597 and 587/586 windows, transport loss, selection and reproduction. [J03](https://www.degruyterbrill.com/document/doi/10.1515/9781575065533/html), [J04](https://www.eisenbrauns.org/books/titles/978-1-934309-57-5.html), [J05](https://doi.org/10.1163/9789004365421) |
| E05 · Return and continuing eastern diaspora (MES → LEV_S; evidence B) | Babylonian residents persist across the Persian period. | Retain outward and return routes alongside a diaspora reservoir. | Return permission is not a count; no wholesale return assumption. [J03](https://www.degruyterbrill.com/document/doi/10.1515/9781575065533/html), [J05](https://doi.org/10.1163/9789004365421) |
| E06 · Egypt–Levant settlement channels (LEV_S → EGY; evidence C) | Persian-period Elephantine household evidence is localized; earlier connectivity needs separate extraction. | Soldiers, refugees and families are possible reproductive bridges. | This broad research window is not a measured continuous route or an Exodus chronology. [J06](https://www.jstor.org/stable/j.ctv346sfv) |
| E07 · Ptolemaic Egypt and Alexandria (LEV_S → EGY; evidence B) | Documentary Jewish communities precede AD 70. | A large diaspora gateway may lead into Egyptian and Mediterranean hosts. | Presence and documentary density do not measure intermarriage. [J07](https://cris.huji.ac.il/en/publications/the-ptolemaic-period-323-bce30-bce/), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html) |
| E08 · Anatolia and Aegean urban networks (LEV_N → AAB; evidence B) | Hellenistic/Roman diaspora scholarship supports urban communities. | Serial stepping stones, if families settle and cross boundaries. | Extract city-specific first attestations; no instant mixing across Anatolia. [J08](https://www.cambridge.org/core/books/abs/cambridge-history-of-judaism/diaspora-in-the-roman-period-before-ce-70/AD1E5333322627A5BD5063DB966A0813), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html) |
| E09 · Eastern Mediterranean to Rome/Italy (AAB → ITM; evidence B) | Western communities predate 70; Rome has wider Mediterranean reproductive connections. | Potential access to Italian host networks. | Ancient DNA does not identify Jewish migrants; distinguish slavery, manumission and free settlement. [J08](https://www.cambridge.org/core/books/abs/cambridge-history-of-judaism/diaspora-in-the-roman-period-before-ce-70/AD1E5333322627A5BD5063DB966A0813), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html), [G15](https://pmc.ncbi.nlm.nih.gov/articles/PMC7093155/) |
| E10 · Egypt, Cyrenaica and North African routes (EGY → MAG; evidence C) | Jewish diaspora scholarship documents eastern North African communities and the 115–117 rupture. | Gateway can contract after opening; survivors elsewhere may preserve branches. | Cyrenaica must be separate from Maghreb before imposing the revolt shock. [J07](https://cris.huji.ac.il/en/publications/the-ptolemaic-period-323-bce30-bce/), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html) |
| E11 · Roman captivity and relocation (LEV_S → ITM; evidence C) | Wars generated destruction and displacement; detailed destinations and totals need extraction. | Deaths and slavery can extinguish or transport branches. | No total expulsion; localize Judean losses and separate the three war windows. [J08](https://www.cambridge.org/core/books/abs/cambridge-history-of-judaism/diaspora-in-the-roman-period-before-ce-70/AD1E5333322627A5BD5063DB966A0813), [J14](https://www.cambridge.org/core/journals/journal-of-roman-archaeology/article/cassius-dios-figures-for-the-demographic-consequences-of-the-bar-kokhba-war-exaggeration-or-reliable-account/2D9838573C499EFC816F5E1569E63B58) |
| E12 · Conversion, incorporation and affiliation exits (LEV_S → LEV_S; evidence C) | Conversion and institutional boundaries are debated and regionally variable. | Social reassignment changes mate access without changing the descent tag. | Conversion alone never creates descent; legal bans do not quantify behavior. [J06](https://www.jstor.org/stable/j.ctv346sfv), [J10](https://academic.oup.com/book/4777), [J11](https://www.degruyterbrill.com/document/doi/10.1515/9781400824854/html), [J15](https://academic.oup.com/book/1791) |
| E13 · Babylonian–Iranian and frontier networks (MES → IRC; evidence B) | Jewish society participated in Sasanian imperial institutions. | Potential host entry and eastward travel require separate gates. | Participation and scholarly travel do not establish host reproduction. [J12](https://www.cambridge.org/core/books/babylonian-jews-and-sasanian-imperialism-in-late-antiquity/BFEE31AF0F2AC14877F81E4D7CBEA339) |
| E14 · Northwestern Arabian oasis networks (LEV_S → ARB; evidence C) | Late-antique communities have uneven epigraphic and literary support. | Possible entry to caravan and later host networks. | Origins, arrival windows and settler shares unresolved. [J13](https://doi.org/10.1017/9781139048873.013) |
| E15 · Himyar and Red Sea interfaces (ARB → HEA; evidence C) | Late-antique Judaism and political adoption in Himyar are supported. | Test immigrant-associated and primarily local-conversion histories. | State religion supplies no founder ancestry; coastal reproduction needs direct evidence. [J13](https://doi.org/10.1017/9781139048873.013) |
| E16 · Northern Levant and Mesopotamian corridor (LEV_N → MES; evidence C) | Imperial resettlement and later frontier societies provide route context. | Separate upper/lower Mesopotamia and port/interior towns. | Magnitude and successful household crossings await targeted extraction. [J02](https://epub.ub.uni-muenchen.de/58863/), [J12](https://www.cambridge.org/core/books/babylonian-jews-and-sasanian-imperialism-in-late-antiquity/BFEE31AF0F2AC14877F81E4D7CBEA339) |
| E17 · Diaspora enclave to host households (ITM → ITM; evidence C) | Marriage and community evidence vary across ancient and later diaspora settings. | First host child and later host participation can matter more than migrant volume. | This is a template repeated per locality, not one Italy-wide or all-Jewish rate. [J06](https://www.jstor.org/stable/j.ctv346sfv), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html), [G16](https://doi.org/10.1016/j.cell.2022.11.002) |
| E18 · Mediterranean hosts to European hinterlands (ITM → WNE; evidence C) | Regional movement and medieval population heterogeneity motivate differentiated networks. | Assimilated descendants follow host mobility into later emigrant sources. | Research locality-specific connections; no homogeneous Europe. [G15](https://pmc.ncbi.nlm.nih.gov/articles/PMC7093155/), [G16](https://doi.org/10.1016/j.cell.2022.11.002) |
| E19 · European Jewish settlement and displacement (WNE → ECE; evidence C) | Medieval Jewish heterogeneity and founder history require internal differentiation. | Enclave persistence, conversion and eastward movement create alternative paths. | Expulsions, bottlenecks and union rates need detailed extraction. [G16](https://doi.org/10.1016/j.cell.2022.11.002), [J09](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html) |
| E20 · Gulf/Red Sea to South Asia (ARB → SAS; evidence C) | Indian Ocean commerce and South Asian social substructure are documented. | Port households may bridge networks while caste/community boundaries slow diffusion. | Trade volume does not supply reproductive rates; split coastal and inland strata. [G07](https://www.cambridge.org/core/books/trade-and-civilisation-in-the-indian-ocean/05F37912D2E7BC563B8D88F4A1FABD03), [G14](https://pmc.ncbi.nlm.nih.gov/articles/PMC3769933/) |
| E21 · Iranian and Central Asian corridors (IRC → CAS; evidence D) | Regional route context requires corridor-specific historical extraction. | Possible eastward gateway once source contains descendants. | Candidate research edge, not confirmed founder or quantified reproductive route. [G07](https://www.cambridge.org/core/books/trade-and-civilisation-in-the-indian-ocean/05F37912D2E7BC563B8D88F4A1FABD03) |
| E22 · Central Asian to East Asian gateways (CAS → EAM; evidence D) | Rohde supplies a coarse benchmark connection, not calibration. | Tests a potentially limiting eastern gateway. | Acquire independent corridor and household evidence before numeric use. [R_B](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementB.pdf) |
| E23 · Indian Ocean coastal reproductive exchange (ARB → HEA; evidence B) | Medieval coastal genomes support African–Asian reproduction and sex bias. | A real reproductive proxy strengthens route plausibility. | Sampled coast/elite contexts cannot be extended to all inland populations. [G08](https://www.nature.com/articles/s41586-023-05754-w), [G07](https://www.cambridge.org/core/books/trade-and-civilisation-in-the-indian-ocean/05F37912D2E7BC563B8D88F4A1FABD03) |
| E24 · Madagascar founding sources (ISEA → MAD; evidence B) | Genomic evidence supports multiple island founding sources. | A rare long-distance reproductive bridge. | Refine arrival timing and ensure source lineage is present before departure. [G09](https://pmc.ncbi.nlm.nih.gov/articles/PMC5559028/) |
| E25 · Pre-European Pacific–American contact (OCE → SAM; evidence B) | Modern and ancient genomes support localized pre-European contact. | Alternative topology can change entry timing. | Direction, island/source locality and counts need scenario treatment; no entire-continent coverage. [G10](https://doi.org/10.1038/s41586-020-2487-2), [G11](https://www.nature.com/articles/s41586-024-07881-4) |
| E26 · Atlantic colonization and settlement (IBE → MCA; evidence B) | Oceanic migration and subsequent reproduction reshaped receiving populations. | Large new gateway plus separate Indigenous/settler social boundaries. | MCA is only a container; each coast/interior/island route requires its own evidence. [G12](https://legacy.slavevoyages.org/blog/methodology-trans-atlantic), [G13](https://pmc.ncbi.nlm.nih.gov/articles/PMC7413858/) |
| E27 · Forced Atlantic migration (SWA → MCA; evidence A) | Voyage records constrain embarked and arrived people, with reconstruction uncertainty. | Directed movement plus survival and highly unequal reproduction. | Do not count arrivals as mixed unions; disaggregate African sources and American destinations. [G12](https://legacy.slavevoyages.org/blog/methodology-trans-atlantic), [G13](https://pmc.ncbi.nlm.nih.gov/articles/PMC7413858/) |
| E28 · Atlantic settlement and mass migration (WNE → NAM; evidence B) | Later modern origin/destination data constrain movement; earlier records need extraction. | Connect European host lineages to new receiving networks. | Preserve Indigenous, immigrant and religious strata. [G05](https://www.un.org/development/desa/pd/content/international-migrant-stock), [G06](https://doi.org/10.1038/s41597-022-01271-z) |
| E29 · Australia and Pacific settlement corridors (WNE → AUS; evidence C) | Modern movement can be checked with migration data; early local histories remain to extract. | Late routes leave few generations for onward spread. | Settler entry does not imply reproductive access to all Indigenous communities. [G05](https://www.un.org/development/desa/pd/content/international-migrant-stock), [G06](https://doi.org/10.1038/s41597-022-01271-z) |
| E30 · South Asian to maritime Southeast Asian gateways (SAS → ISEA; evidence C) | Indian Ocean institutional routes motivate targeted port research. | Connect settled coastal families to island chains. | No immediate access to all inland or island mating pools. [G07](https://www.cambridge.org/core/books/trade-and-civilisation-in-the-indian-ocean/05F37912D2E7BC563B8D88F4A1FABD03), [G14](https://pmc.ncbi.nlm.nih.gov/articles/PMC3769933/) |
| E31 · Island chains to Near Oceania (ISEA → PNG; evidence D) | Benchmark island connectivity needs modern locality-specific evidence. | Island substructure controls stepwise spread. | Retain highland/coastal barriers; this research window is not an activation date. [R_B](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementB.pdf) |
| E32 · Near to Remote Oceania (PNG → OCE; evidence D) | Settlement/contact sequences require island-level reconstruction. | Edges may predate arrival of founder descendants at the source. | A route existing early is not useful unless descendants reach it before it closes. [R_B](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementB.pdf), [G11](https://www.nature.com/articles/s41586-024-07881-4) |

The initial register deliberately does not claim complete connectivity for every node. Next extraction must add locality-specific links (including North Africa–Sahel, East Asia–Korea/Japan, East Asian/Southeast Asian coasts, African interior gateways, intra-American connections and contemporary receiving populations) when route tests make them consequential. Run no historical forecast from this sparse skeleton with absent links silently set to zero.

Boundary-crossing groups receive focused research: settled merchants, military households, captives, enslaved and freed people, refugees, converts and mixed households. Elite or royal marriages can demonstrate permeability but cannot supply ordinary-person rates. For each category ask who remained, who had surviving children, with whom, and in which social network those children participated.

## 6. Historical Adaptive social-state model

Use states as local, time-indexed participation categories, separate from immutable genealogy:

`core community ↔ diaspora enclave ↔ mixed/boundary household ↔ transitional participation ↔ host participation`

This is a possible transition graph, not a required one-way ladder. Residence and participation are distinct dimensions: a host-affiliated person can move again; a mixed household can retain enclave ties; siblings can enter different marriage networks. Conditions may allow return/reaffiliation. Assign each state to a specified locality, regime and household context.

| Transition | Evidence/proxy to seek | Consequence for the pedigree process |
|---|---|---|
| Core to enclave | Residence, deportation, military or family records | Change town and location; keep enclave partner preferences until evidence or a scenario changes them. |
| Enclave to boundary household | Marriage/household records, cautiously interpreted admixture | Record actual parents and a cross-group child; a migrant alone does not trigger it. |
| Boundary to transitional participation | Children’s partners, institutions, names with corroboration | Modify partner access gradually or with a scenario-specific lag. |
| Transitional to host participation | Host institutions and subsequent marriage patterns | Use current host migration and mating behavior; keep the descent flag unchanged. |
| Conversion or reaffiliation | Local legal, narrative and household evidence | Change participation when appropriate; never add or erase biological descent by conversion alone. |

Do not permanently assign “Jewish migration” to descendants who have joined a different social population. Do not double-penalize cross-group unions by applying both endogamy and an independent barrier coefficient that encode the same phenomenon. Fit a single partner-choice matrix or clearly conditional mechanism. Rare successful crossings may have large later effects even when the average flow is small.

## 7. Machine-readable contract

Deliverables include `historical-architecture.json` and `historical-model.schema.json`. The first is a research registry with intentionally unfilled numerical fields. The schema supports later evidence-backed parameter records. Unknown is `null`, never zero. The future production runner must reject unfilled required parameters instead of silently inventing defaults.

Use astronomical years internally: 1800 BCE = −1799; 1 BCE = 0; AD 1 = 1. Intervals include their start and exclude their end. Human-facing labels use BCE/AD. Simulation horizon is AD 2026; do not inherit the existing exploratory simulator’s AD 3000 horizon.

| Record | Required content and units |
|---|---|
| Geographic/local node | Stable ID; name; parent; boundary definition; active interval; geometry version; count scope; local/town hierarchy; population series with persons, uncertainty and observed/reconstructed/interpolated/projected status. |
| Social population | Locality; affiliation/state definition by regime; community as a subset of geographic total; partner-choice matrix; urban/rural or class distinction only where consequential. |
| Demography | Age/sex survival and reproduction; adult-only versus all-age convention; offspring variance/childlessness; partnership/remarriage; local mating strength; population targets and event shocks. |
| Directed edge | Source/destination locality and social state; date/attestation intervals; departure/arrival/settled/reproductive-settler measure; event total or per-generation flow with denominator; burst duration; transport/distance; household clustering and age/sex selection. |
| Conditional stages | Settlement given arrival; reproduction given settlement; cross-group parenthood given reproduction; social-state transition/delay. Explicitly state when these are summaries of individual events rather than extra multipliers. |
| Parameter evidence | Historical evidence; inference; numerical choice/distribution; units/time basis/denominator; separate evidence and numerical grades; source IDs and page/table/record locator; dependency group; scenario; status and last review date. |
| Run manifest | Engine/data versions and hashes; seed; historical parameter draw; pedigree replicate; founder condition; all inputs; population scale; time convention; outcomes and censoring indicators. |

Keep geography, social-state transitions, population anchors, shocks, parameters, evidence and scenario dependencies in separate linked tables. Avoid one giant row per country-century. Event flows should conserve people across origin/destination and distinguish transit deaths. Do not clone tagged individuals to populate new colonies or duplicate nested Jewish/host totals. Migration stocks, migration flows and genetic effective population size have different denominators.

## 8. Historical population backbone

Use global population scaffolding at approximately century points for integration only where helpful, retaining the original evidence’s dates. HYDE 3.2 offers a broad allocation framework, but original BCE outputs are millennium-spaced and a fine grid is not a local observation. Mark generated century anchors as interpolation; retain source version and interpolation rule. Ancient regional settlement surveys should replace this scaffold where they offer better geographically matched bounds. [HYDE](https://doi.org/10.5194/essd-9-927-2017).

Audit older totals instead of layering false precision on them. McEvedy–Jones is useful as a comparison and provenance trail, not ground truth; Maddison and other compilations may share underlying estimates and therefore fail to provide independent corroboration. [Guinnane (2023)](https://doi.org/10.1017/S0022050723000293), [Maddison Project](https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2023?lang=en).

Use UN WPP for modern age/sex, fertility, survival and population inputs; retain the vintage and estimate/projection distinction. AD 2026 is a projection in WPP 2024, not a measured census total. Modern migration stocks and modeled flow estimates help with transport/residence but need independent marriage and fertility evidence. [UN WPP](https://population.un.org/wpp/), [UN Migrant Stock](https://www.un.org/development/desa/pd/content/international-migrant-stock), [Abel & Cohen](https://doi.org/10.1038/s41597-022-01271-z).

The first population extraction should cover Judah/Samaria/Galilee, upper/lower Mesopotamia and Egyptian gateway settlements. Retain competing population reconstructions, definitions of occupied settlement area and density assumptions. No ancient population values are filled merely to make the graph executable.

## 9. Evidence grades and uncertainty

- **A:** direct demographic/historical evidence or strong independent convergence for the particular claim.
- **B:** reasonably supported but incomplete.
- **C:** indirect proxy or substantial dispute.
- **D:** model choice without adequate direct historical calibration.

Grade each field. A document may strongly establish a family’s residence while providing no denominator for a settlement or intermarriage rate. Record whether a source was read in full, only selected passages were inspected, or the bibliography was verified from its record/abstract. Publisher metadata alone does not constitute a completed historical extraction.

Use bounded alternatives and only specify a probability distribution when justified. Low/central/high are not automatically quantiles; the central interpretation is not a historical posterior. Separate graph uncertainty, date uncertainty, demographic uncertainty and chance events within a pedigree. Correlate related parameters so that migration totals, diaspora growth and population targets do not contradict each other.

Genetic evidence constrains reproductive contact, timing, sex bias and structure. It never directly supplies the chance of descent from this founder. Absence of detectable DNA also cannot prove absence of genealogical descent. Sampled cemetery or elite contexts do not define the entire host’s marriage behavior.

## 10. Scenario architecture

**Historically conservative.** Use securely evidenced settlement paths; test long enclave persistence and late host access. Uncertain routes remain explicit alternatives, not asserted historical zeros.

**Central working interpretation.** Combine well-supported route chronology with heterogeneous communities and occasional boundary crossings. It is a working interpretation, not a measured midpoint or maximum-likelihood history.

**Higher connectivity.** Add plausible weakly documented bridges, earlier feasible openings and faster host participation where not contradicted by evidence. No automatic panmixia or guaranteed survival.

Retain disagreements as explicit axes: local continuity versus stronger collapse; voluntary movement versus different forced-migration contributions; longer enclave persistence versus more porous affiliation; limited versus wider conversion/incorporation; localized versus broader Roman-war impact; and Arabian immigrant-associated versus largely local-conversion histories. These axes can interact; do not independently choose every low value and call it a coherent conservative history. Do not assign equal posterior weights without justification.

## 11. Sensitivity priorities and research order

The following is an initial hypothesis ranking, not a measured historical result:

1. **Founder survival:** Early reproductive failure, correlated household loss, offspring variance. Unconditional survival can dominate; later routes cannot revive an extinct lineage.
2. **First host reproductive crossing:** Local Levantine and Jewish–host unions, exits, settlement and reproduction. Opens a much larger pedigree network; a surviving enclave may otherwise remain enclosed.
3. **Isolation and port timing:** Route existence, direction, opening/closure windows. Time-respecting paths and missed expansion windows matter more than travel totals.
4. **Host participation and local structure:** Enclave clustering, transitional households, host marriage rules. Determines whether an initial bridge spreads beyond a few related households.
5. **Gateway bottlenecks:** War, deportation selection, urban sinks, family migration. Can remove or amplify rare tagged branches; census size alone is insufficient.
6. **Host networks and remote boundaries:** Caste/religious/class barriers, coast/interior and island structure. Often crucial for unreached groups even if world-weighted coverage barely changes.
7. **Demographic refinement:** Age schedules, fertility, cousin marriage and reproductive inequality. Refine when measured sensitivity elevates them; avoid unsupported ancient precision.

Use a nested ensemble: an outer draw selects a coherent historical graph/parameter interpretation; repeated inner simulations sample births, deaths, partner choice and migration. Retain stochastic variation separately from historical uncertainty. Begin with grouped screening and topology experiments; follow with interaction-aware analysis of influential parameters. Independent Sobol indices are inappropriate when the parameters are correlated without accounting for that dependence.

Test route deletion, later openings, equal total migration with different bursts, clustered versus dispersed settlement, different host-entry lags, and geography splits with conserved populations. Use reproducible seed lists; paired seeds help only to the extent event paths remain comparable. Record Monte Carlo error. Zero observed crossings in a small ensemble does not prove zero probability. A parameter can affect unconditional survival, arrival time and final regional coverage differently; publish separate rankings.

**Ordered research passes and exit criteria**

1. **Architecture and mechanical validation:** explicit founder, unit conventions, tag neutrality, extinction, chronological ordering, local mating and directed ports. Compare tiny pedigrees against independent ancestor traversal. The supplied synthetic diagnostic begins this pass.
2. **First evidence extraction:** household/boundary evidence for the Levant, Babylonian family archives, Egyptian papyri and Mediterranean enclaves. Extract locators, sample definitions and what each source cannot measure. Record null if no representative rate exists.
3. **First historical pilot, approximately 1800 BCE–AD 700:** use finitely represented local populations with transparent demographic bounds, core/enclave/host participation and documented movement scenarios. Include a Rohde-compatibility benchmark; do not claim Supplement B replication from graph tests.
4. **Sensitivity-directed receiving networks:** refine the Mediterranean and eastern/Arabian gateways that change first passage and host entry. Increase South Asian/social and coast/interior detail before interpreting broad coverage.
5. **Later global network:** research the Atlantic, Indian Ocean and Pacific routes the pilot can actually reach; refine Indigenous and residual barriers before any universality output.
6. **Population and demographic refinement:** use new evidence only where it changes establishment, extinction or materially different coverage. Test scale, number of towns and alternative preserved-flow conventions. An unvalidated small-scale run is not a scaled historical forecast.
7. **Release gate:** a quantitative Historical Adaptive historical result requires a versioned complete parameter set, provenance for each assumption, explicit topology alternatives, calibrated population accounting, the full stage-separated process, convergence checks and nested uncertainty runs. Publish remaining non-identifiability with the result.

Research priority should depend on expected reduction in decision-relevant uncertainty per effort. A sensitive but unidentifiable ancient rate stays a scenario uncertainty; repeated searching does not turn it into evidence.

## 12. Parameters for which good historical rates probably do not exist

| Uncertain quantity | What evidence might constrain | What must remain an assumption |
|---|---|---|
| Founder survival/reproductive output | General demographic context | This person’s realized pedigree, age/household and surviving children. |
| Ordinary ancient intermarriage | Marriage contracts, named households, cautious genomic admixture | Representative group-wide percentages without denominators. |
| Permanent settlement of travelers | Repeated household attestations, cemeteries, landholding | Fraction of all travelers who settled. |
| Reproduction of migrants/captives | Family documents, descendants, sex/age structure | Population-wide lifetime reproduction probabilities. |
| Assimilation duration | Children’s marriages and institutional participation | A single universal number of generations. |
| Ancient annual migration | Event narratives, archaeological movements and settlement | Continuous annual flow, its age/sex selection and unrecorded tail. |
| Local mate pool/kin preference | Site sizes, mobility proxies, occasional kinship data | Exact town-scale choice weights over millennia. |
| Rare undocumented bridges | Evidence that a boundary was permeable | Their exact dates/counts; non-documentation is not zero. |
| Household-correlated catastrophe loss | Settlement destruction and selective narratives | Fine family survival correlations and precise mortality totals. |

Use model-assumption ranges for these quantities only in explicitly labeled experiments. Do not turn a genetic ancestry percentage into any of them.

## 13. Initial annotated bibliography

These sources are starting points for targeted extraction, not a claim that all books have been fully reviewed. The linked machine-readable ledger records access status. The regional detail above is a synthesis/research inference; numerical parameters remain unestimated.

- **R_2004 — [Rohde, Olson & Chang (2004), Modelling the recent common ancestry of all living humans. Nature 431, 562–566.](https://doi.org/10.1038/nature02842)** Foundational structured genealogy; distinguish a global MRCA from a specified founder.
- **R_A — [Rohde, Olson & Chang (2004), Supplementary Methods A](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementA.pdf)** Graph-model assumptions and mathematical checks, not historical rates.
- **R_B — [Rohde, Olson & Chang (2004), Supplementary Methods B](https://www.stat.yale.edu/~jtc5/papers/CommonAncestors/NatureCommonAncestors-SupplementB.pdf)** Nested towns/countries/regions, lifetimes, clustering, dated ports and scaling limits.
- **J01 — [Agranat-Tamir et al. (2020), The Genomic History of the Bronze Age Southern Levant. Cell.](https://pubmed.ncbi.nlm.nih.gov/32470400/)** Movement and admixture constraints; cannot identify Jacob or measure intermarriage rates.
- **J02 — [Radner (2018/2019), The “Lost Tribes of Israel” in the Context of the Resettlement Programme of the Assyrian Empire.](https://epub.ub.uni-muenchen.de/58863/)** Documented Assyrian destinations and uncertain extensions; no modern lost-tribe identifications.
- **J03 — [Lipschits (2005), The Fall and Rise of Jerusalem: Judah under Babylonian Rule.](https://www.degruyterbrill.com/document/doi/10.1515/9781575065533/html)** Regional destruction and continuity; compare alternative archaeological reconstructions.
- **J04 — [Pearce & Wunsch (2014), Documents of Judean Exiles and West Semites in Babylonia. CUSAS 28.](https://www.eisenbrauns.org/books/titles/978-1-934309-57-5.html)** Resident families in primary-document editions; selected archive, not a population denominator.
- **J05 — [Alstola (2020), Judeans in Babylonia: A Study of Deportees in the Sixth and Fifth Centuries BCE.](https://doi.org/10.1163/9789004365421)** Rural/urban settlement, social position and differential integration.
- **J06 — [Satlow (2001), Jewish Marriage in Antiquity.](https://www.jstor.org/stable/j.ctv346sfv)** Household evidence and variation; distinguish prescribed marriage rules from practice.
- **J07 — [Hacham & Ilan, eds. (2020), Corpus Papyrorum Judaicarum IV: The Ptolemaic Period.](https://cris.huji.ac.il/en/publications/the-ptolemaic-period-323-bce30-bce/)** Egyptian documentary evidence; use with CPJ V (2022), Early-Roman Period, 30 BCE–117 CE.
- **J08 — [Smallwood (1999), The Diaspora in the Roman Period before CE 70. Cambridge History of Judaism III.](https://www.cambridge.org/core/books/abs/cambridge-history-of-judaism/diaspora-in-the-roman-period-before-ce-70/AD1E5333322627A5BD5063DB966A0813)** Pre-70 diaspora chronology and voluntary as well as forced movement.
- **J09 — [Gruen (2002; 2004 edition linked), Diaspora: Jews Amidst Greeks and Romans.](https://www.degruyterbrill.com/document/doi/10.4159/9780674037991/html)** Mediterranean communities and participation; cultural integration is not a measured mating rate.
- **J10 — [Goodman (1994), Mission and Conversion.](https://academic.oup.com/book/4777)** Conversion debate; retrieve opposing interpretations before calibration.
- **J11 — [Schwartz (2001), Imperialism and Jewish Society, 200 BCE to 640 CE.](https://www.degruyterbrill.com/document/doi/10.1515/9781400824854/html)** Interpretive social reconstruction for alternative boundary regimes, not consensus rates.
- **J12 — [Gross (2024), Babylonian Jews and Sasanian Imperialism in Late Antiquity.](https://www.cambridge.org/core/books/babylonian-jews-and-sasanian-imperialism-in-late-antiquity/BFEE31AF0F2AC14877F81E4D7CBEA339)** Differentiate Babylonian Jewish society within imperial institutions.
- **J13 — [Robin & Harris (2021), Judaism in Pre-Islamic Arabia.](https://doi.org/10.1017/9781139048873.013)** Distinguish northwestern oases, Himyar, local conversion and settler origins.
- **J14 — [Raviv & Ben David (2021), Cassius Dio’s figures for the demographic consequences of the Bar Kokhba War: exaggeration or reliable account?](https://www.cambridge.org/core/journals/journal-of-roman-archaeology/article/cassius-dios-figures-for-the-demographic-consequences-of-the-bar-kokhba-war-exaggeration-or-reliable-account/2D9838573C499EFC816F5E1569E63B58)** A catastrophe reconstruction to test; do not turn literary totals into exact mortality.
- **J15 — [Lapin (2012), Rabbis as Romans: The Rabbinic Movement in Palestine, 100–400 CE.](https://academic.oup.com/book/1791)** Avoid uniform or anachronistic rabbinic boundaries.
- **G01 — [Klein Goldewijk et al. (2017), HYDE 3.2. Earth System Science Data.](https://doi.org/10.5194/essd-9-927-2017)** Broad population scaffold; fine grids are allocations. Ancient century values would be interpolations.
- **G02 — [Guinnane (2023), We Do Not Know the Population of Every Country in the World for the Past Two Thousand Years. Journal of Economic History.](https://doi.org/10.1017/S0022050723000293)** Audit weak ancient totals and unjustified disaggregation of McEvedy–Jones estimates.
- **G03 — [Maddison Project Database 2023](https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2023?lang=en)** Cross-check source sheets; shared underlying estimates are not independent evidence.
- **G04 — [UN World Population Prospects 2024](https://population.un.org/wpp/)** Modern population/age/sex/fertility; 2026 is a projection in this vintage.
- **G05 — [UN International Migrant Stock 2024](https://www.un.org/development/desa/pd/content/international-migrant-stock)** Modern residence stocks are not migration flows or intermarriage counts.
- **G06 — [Abel & Cohen (2022), Bilateral international migration flow estimates updated and refined by sex.](https://doi.org/10.1038/s41597-022-01271-z)** Modeled flow estimates still need separate settlement and reproductive evidence.
- **G07 — [Chaudhuri (1985), Trade and Civilisation in the Indian Ocean.](https://www.cambridge.org/core/books/trade-and-civilisation-in-the-indian-ocean/05F37912D2E7BC563B8D88F4A1FABD03)** Corridors and institutions, not measured reproduction.
- **G08 — [Brielle et al. (2023), Entwined African and Asian genetic roots of medieval peoples of the Swahili coast.](https://www.nature.com/articles/s41586-023-05754-w)** Reproductive exchange and sex bias; coastal samples do not measure inland diffusion.
- **G09 — [Pierron et al. (2017), Genomic landscape of human diversity across Madagascar.](https://pmc.ncbi.nlm.nih.gov/articles/PMC5559028/)** Island founding sources and substructure; not a founder-specific ancestry claim.
- **G10 — [Ioannidis et al. (2020), Native American gene flow into Polynesia predating Easter Island settlement.](https://doi.org/10.1038/s41586-020-2487-2)** Alternative pre-European Pacific–American reproductive topology.
- **G11 — [Moreno-Mayar et al. (2024), Ancient Rapanui genomes reveal resilience and pre-European contact with the Americas.](https://www.nature.com/articles/s41586-024-07881-4)** Constrain timing of a localized Pacific–American bridge, not continental mixing.
- **G12 — [SlaveVoyages, Trans-Atlantic methodology](https://legacy.slavevoyages.org/blog/methodology-trans-atlantic)** Embarked/arrived distinctions and missing voyages; onward reproduction is separate.
- **G13 — [Micheletti et al. (2020), Genetic Consequences of the Transatlantic Slave Trade in the Americas.](https://pmc.ncbi.nlm.nih.gov/articles/PMC7413858/)** Sex-biased reproductive histories cannot be inferred solely from shipping totals.
- **G14 — [Moorjani et al. (2013), Genetic Evidence for Recent Population Mixture in India.](https://pmc.ncbi.nlm.nih.gov/articles/PMC3769933/)** Retain social/endogamous substructure inside South Asia.
- **G15 — [Antonio et al. (2019), Ancient Rome: A genetic crossroads of Europe and the Mediterranean.](https://pmc.ncbi.nlm.nih.gov/articles/PMC7093155/)** Mediterranean reproductive connections in sampled Rome; no Jewish-specific rate.
- **G16 — [Waldman et al. (2022), Genome-wide data from medieval German Jews show that the Ashkenazi founder event pre-dated the 14th century.](https://doi.org/10.1016/j.cell.2022.11.002)** Medieval heterogeneity and bottlenecks; effective population size is not census size.

## Deliverable boundary

Completed in this release: research architecture, proposed geographic/regime/edge registries, evidence separation, schema, alternative scenarios, ranked research plan and a small reproducible mechanics diagnostic. Not completed: calibrated ancient population series, a fully evidenced numerical connectivity graph, nested 1800 BCE–2026 historical Monte Carlo ensembles, or a historical estimate of modern Jacob-descendant coverage. The new Site page keeps those limits visible while providing the working architecture for the next research pass.