Ep. 43: Crossing the U.S. Border - Hispanic Roots and AI

Three AI tools, one 1866 Mexican parish register, and the family 
that a U.S. census reduced to a single word: México.

Brian traces composite ancestor Esteban Vasquez through two Arizona 
census records where birthplace reads only "México" and no 
naturalization record exists. The research pivots to FamilySearch's 
Mexico-Sonora Catholic Church records, a collection covering roughly 
21 specific parishes from the mid-1600s onward, and finds the 1866 
baptismal entry that names his parents and the mother's maiden 
surname no American record ever recorded.

Three tools, each matched to the job: Perplexity and Comet map the 
Mexican archive landscape before the search begins. Gemini via AI 
Studio transcribes a handwritten Spanish-language baptismal record. 
Claude correlates the transcription against census data and 
identifies the parents, the two-surname naming convention, and the 
compadrazgo (godparent) network as the next research thread.

Copy-paste prompts for all three tools. Clear explanations of 
FamilySearch's Sonora Catholic records, the Registro Civil, and how 
to navigate Mexican archives when your census says only "born Mexico."

This episode is for you if you search: AI tools for Hispanic 
genealogy, FamilySearch Mexico Catholic records, Sonora parish 
registers, Spanish handwriting transcription AI, how to find Mexican 
ancestors, Registro Civil research, two-surname genealogy research.

For Australian and UK listeners: the same three-tool workflow applies 
to Catholic parish research in Ireland, England, and Scotland through 
Findmypast's Catholic Heritage Archive and the National Library of 
Ireland at registers.nli.ie.

Full Breakthrough. The 1866 baptismal record names the parents. The 
research moves forward.

Advanced prompts and the full Companion Guide at ancestorsandai.com.

Connect with Ancestors and Algorithms:

📧 Email: ancestorsandai@gmail.com
🌐 Website: https://ancestorsandai.com/
📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/

Golden Rule Reminder: AI is your research assistant, not your researcher.

Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research!

New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

Transcript

SPEAKER_00 0:12

Three words. That's all a census form said about where he came from. Mexico. No state, no city, no village. Just Mexico, written in the neat hand of a census enumerator who either didn't know or didn't think it mattered. For tens of millions of Americans with Hispanic heritage, that wall is familiar. The paper trail crosses the border and stops. The ancestors on the other side feel unreachable. Until now. Today I'm going to show you how AI broke that wall wide open using tools you can access for free on records that go back centuries. And I mean centuries, not decades. This is Ancestors and Algorithms, and this might be the episode some of you have been waiting a very long time for. Let's dive in. Welcome to Ancestors and Algorithms, where family history meets artificial intelligence. I'm your host Brian, and today we're going somewhere we've never gone on this show before. We're crossing the border into Mexican genealogy, and I want to show you that the research world on the other side is extraordinary. If your family has roots in Mexico, in the American Southwest, or anywhere in Latin America, this episode is for you. And if they don't, stick with me, because the methods we're covering today, AI powered transcription of foreign language records and navigating a major international archive system, apply no matter where your ancestors lived. So let's get started. I want to tell you about Esteban. Full name Esteban Vasquez Ibarra. Born sometime around 1866, somewhere in the state of Sonora, Mexico. Died in Tucson, Arizona, probably in the late 1920s, and for the better part of a century, his family knew almost nothing about where he came from before he arrived in the Southwest. Esteban is a composite ancestor, a case I've built to illustrate one of the most common brick walls in Hispanic genealogy research. But the type of mystery he represents is one I hear about constantly. The details change. The frustration is always the same. Here's what the family had when they started. Esteban on the 1900 census in Pima County, Arizona, age listed as 35, birthplace Mexico, wife Carmen, Arizona born, children born in Arizona, and then Esteban again on the 1910 census, same county, a few years older, occupation listed as labor, birthplace, once again Mexico. No state, no city, no village, just the enormous country that had been his home and was now his past. Now I want to stop here for a moment because something matters that I think deserves more than a footnote. The border between the United States and Mexico has never been a simple line. After the Treaty of Guadalupe Hildago in eighteen forty eight and then the Gadsden Purchase in 1853, roughly half of Mexico became the American Southwest. Families who had lived their entire lives on Mexican land suddenly found themselves in a different country without moving an inch. The promises made in that treaty about land rights and citizenship were not always honored. Communities that existed for generations were suddenly on the wrong side of a new boundary, and the records that documented their lives were now to different nations' archives. For the generations that came after, men like Esteban who crossed their border looking for work for opportunity for a different life, the documentary trail on either side can feel like it belongs in two different worlds. It doesn't. And today I want to show you how to bridge them. I also want to acknowledge something before we go any further. For many families, this research carries real weight. The history of Mexican Americans in the Southwest, the disruptions of the border, the generations who were told their heritage was something to set aside rather than something to research, that history isn't abstract. If this is your family story, I understand why the search can feel difficult. And I believe what we find on the other side of that difficulty is worth looking for. So back to Esteban. His family's working assumption, which is nearly universal in these situations, was that a naturalization record held the key. If Esteban had become a US citizen, there'd be a declaration of intention, a petition for naturalization, maybe a citizenship certificate. Those records typically included the applicant's exact place of birth. Problem solved. I know some of you are nodding right now because you've had exactly that thought about your own ancestor. The naturalization record is going to tell me everything I need to know. That was the first thing we looked for. And that's where the first wall appeared. Before I explained what happened, I want to be clear about my starting approach. AI is your research assistant, not your researcher. And the first thing I needed from any research assistant was a map of the territory, because if I was going to find Esteban, I had to understand what records actually existed and where they lived before I started digging. The mystery I'm pursuing today, and this is the question that drives everything we do, exactly where in Mexico did Esteban Vasquez Ibarra come from and who were his parents? Sounds simple. It's going to take us somewhere I did not expect. The first tool I reached for was perplexity with its comet browser. And here's why I chose it for this first step rather than jumping straight into an archive. When you're working with a record system you haven't used before, the worst mistake you can make is to start searching without understanding what actually exists. Mexican genealogy has a rich, layered record landscape. It's not organized the way American records are. The archives don't always have obvious English language interfaces. And if you go looking for things that don't exist or that haven't been digitized yet, you'll spend hours on dead ends before you figure out why. Perplexity strength is current web research with citations. I can ask it what exists, where it lives, and whether it's accessible online right now, and I get sourced answers drawn from the current web, not a snapshot from whenever a model was last trained. Comet adds agentic browsing when I need it to dig deeper into a specific page or archive interface. Here's exactly what I asked. Quote, I'm researching a Mexican ancestor from approximately the eighteen sixties who settled in Arizona by the eighteen nineties. I need to understand the genealogy record landscape for Mexico. Please explain one, what Catholic parish records exist in Mexico and where are they typically held? How far back do they go? Two, what is the Registro of Civil and when did it begin? What records does it include? three, what collections does Family Search currently have for Mexican states, particularly Sonora? Four, are there any online Mexican genealogy portals or databases I should know about for current access? Five, what records would typically document a person born in Sonora in the eighteen sixties? Please include sources for your answers, end quote. What came back was one of the most useful research orientation sessions I've had before starting any project. Let me walk you through the key findings, because this is the foundation everything else builds on. First, the Catholic Parish Records. The Catholic Church has kept records in Mexico since the Spanish colonial era, beginning in the 1500s. For many parishes, the surviving records start in the sixteen hundreds and seventeen hundreds. These are baptismal records, marriage records, and burial records, and they can be extraordinarily detailed. We're talking about a record keeping tradition that predates the United States by more than two centuries. For Australian and Irish researchers, this structure is familiar. Catholic parishes organized by diocese keeping sacramental registers in exactly the same way your own ancestors' parishes did. The scale in Mexico is simply larger and older. These records typically live in three places, the diocesan archives for the region, the individual parishes themselves, and critically for our purposes, family search. The Family History Library has microfilmed millions of pages of Mexican Catholic records over several decades. Many are now available online through Family Search's Mexico collection, organized by state, then diocese, then parish. Beyond Family Search, Perplexity flagged other resources worth knowing, the INA, Mexico's National Anthropology and History Institute, which holds an extensive digital archive of historical documents, and state-level genealogical societies in Mexico that maintain their own research guides and help navigate materials that haven't yet made it online. Second, the Registro Sibyl. Mexico's civil registration system was established by the reform laws of 1859 under President Benito Juarez. These laws transformed the recording of births, marriages, and deaths from the church to civil authorities. Family search holds civil registration records for many Mexican states beginning in the eighteen sixties. Third and most relevant for Esteban, Family Search's Sonora collection. The Mexico Sonora Catholic records available through family search span from the mid 1600s through the early 1900s, browsable by parish. The first step is checking whether your family's likely area is among them. For a man born in Sonora in the eighteen sixties before comprehensive civil registration was established in that state, the Catholic parish records are the primary source. This is the research landscape. Perplexity built the map. Now came the walking. Let me be honest with you about something. I assumed, going into this research, that the naturalization record was the pivot point. Find it, get the birthplace, identify the Mexican town, locate the records for that town. Clean, two step process. That is not what happened. A search of the naturalization indexes through NARA and Fold III, the district court records for Pima County and the Ancestry Catalog for Arizona Naturalization Records produce nothing. No Esteban Vasquez. No E Vasquez. No plausible variant that matched. And here's the thing, that's not unusual at all. Many Mexican immigrants in the American Southwest never naturalized. For some, it was the reality of seasonal or circular migration moving back and forth across the border for work. For others, it was the complicated history of how Mexican Americans were treated even when they did pursue citizenship. For others still, they simply didn't see the need or encountered barriers that made the process inaccessible. Whatever Esteban's reason, he never became a U.S. citizen. No naturalization record exists. What I want you to hear when a search comes back empty is this. That result is information. The wall tells you something. It tells you to look somewhere else. Somewhere else, in this case, was the other side of the border. I navigated to Family Search's Mexico Collections for Sonora and browsed the Catholic parish records available for the Sonora River Valley region. When I say browsed, I want to make that concrete for you because it's a different kind of searching than most of us are used to. At FamilySearch.org, you go to the search tab and down to records and then type in the Mexico Sonora. What you find is a list of dioceses and within each diocese the individual parishes. Each parish collection typically covers a span of years and includes separate volumes for baptisms, marriages, and burials. The interface lets you page through images of the original records. This is browsing, not searching by name, and that's an important distinction. Many of these records haven't been indexed yet, which means a name search won't surface them. You have to know the approximate time period and the geographic region of your ancestor to know where to look. And when you find a page worth examining, you need to be able to read it. That's where Gemini via AI Studio becomes extraordinary. Gemini via AI Studio is the show's standard tool for AI handwriting transcription, and the reason is simple. Its document understanding handles historical manuscripts, including Spanish language documents, with a precision I haven't found matched by general purpose AI. I found a collection of baptismal records for a parish in the Sonora Highlands covering the eighteen fifties through the eighteen eighties. I found an entry worth examining closely. And here's exactly what I asked Gemini to do. Quote, please transcribe this image of a Mexican Catholic baptismal record. The document is in Spanish and appears to be from the eighteen sixties. Please provide one, a complete transcription of all legible text preserving the original Spanish exactly as written. Two, for any words you cannot read clearly, indicate with illegible in brackets and your best guess in brackets afterwards. Three, note any abbreviations you recognize and what they likely expand to. Catholic records commonly use Latin abbreviations. Four, include all names, dates, and place names exactly as written with accent marks preserved. Five, after the full transcription, provide a plain language summary of what this record appears to document. The transcription was not perfect. I want to be very clear about that, because no AI transcription of genuinely difficult historical handwriting ever will be. Ink that has bled through from the other side of the page, words written in abbreviations that were second nature to a nineteenth century Catholic priest and opaque to modern readers, variations in letter forms that reflect a scribal tradition rather than any standard would recognize now. Gemini gave me something incomplete in places, flagged with illegible in brackets, and offered its best guesses in brackets. But it was good enough to give me the structure of the record. And the structure of a Catholic baptismal entry is designed to hold specific information, the date, the child's name, the parents' name, the godparents' names, and the parish. That information was legible. And here's what this entry contained. Dated in eighteen sixty six, a male infant named Esteban, baptized in a parish in the Sonoran Highlands, son of Gerardo Vasquez and Conception Ibara, described as Vicinus, meaning residence of this parish, Godfather Tomaz Carrio. Sit with that for a second. A priest in Sonora recorded those names in eighteen sixty six because the church had been recording names in parish books since the Spanish colonial era. He did it because the sacrament of baptism required documentation, because that documentation was the proof of entry into the community of the faithful. He wrote Conception Ibara because she was that child's mother and the church recorded mothers. Thirty four years later, a US Census enumerator wrote Mexico in the birthplace column and moved on. He didn't ask about Esteban's mother. Her name wasn't the state's business. Her surname before she married wasn't the census' concern. The church knew. The church had always known. Now before any of this mattered genealogically, I had to verify it. AI is your research assistant, not your researcher. And a possible match is not a confirmed match. I brought the transcribed baptismal record into Claude along with everything from the US records and asked for something specific. Quote I'm researching Esteban Vasquez Ibarra, born approximately 1865 to 1866 in Mexico, who appears in Arizona Census Records in 1900 and 1910. Below are two sets of information. US Records, 1900 Census, Pima County, Arizona, Esteban Vasquez, age 35, born Mexico, wife Carmen, Arizona born, children born in Arizona, 1910 Census, Pima County, Arizona, Esteban Vasquez, age forty five, born Mexico, occupation laborer, naturalization records, none found in federal or district court indexes. Mexican baptismal record, transcribed from family search, year eighteen sixty six, name Estevan, Father Gerardo Vasquez, Mother Concepcion Ibarra, Location Sonoran Highlands Parish, Godfather Tomascarillo. Please analyze one, what evidence connects these two records to the same person and how strong is that connection? Two, what discrepancies exist and which record would you consider more authoritative for each type of information? Three, what does the baptismal record contribute that the US records cannot and vice versa? Four, what record should I look for next to strengthen or disprove this identification? Five, what can I state with reasonable confidence now versus what still needs verification in quote? Claude's analysis was precise and useful. Let me walk you through it. On the connection, the given name Esteban is exact. The surname Vasquez matches in exactly the way the Mexican naming convention would predict. Under the two surname system, a child's first surname comes from the father's first surname. The father here is Gerardo Vasquez. So Estevan's first surname would be Vasquez. Exact match. The birth year of 1866 aligns with census ages of 35 in 1900 and 45 in 1910, allowing for the well-documented tendency of census ages for immigrants to be approximate. And the geographic logic holds, Sonora is the Mexican state immediately across the border from Pima County, the natural corridor for migration into Tucson. On the discrepancy, the mother surname Ibarra would formerly be Estevant's second surname in the Mexican system, making his full name Esteban Vasquez Ibarra. Yet the U.S. records show him only as Estevon Vasquez. Kalodno to this was entirely expected. The second surname was routinely dropped in American records, sometimes by the immigrant himself for simplicity, sometimes by the record keeper who didn't understand the convention. The absence of Ibarra in the U.S. records isn't a contradiction. It's a predictable artifact of how Mexican names traveled across the border. On what each record contributes, the census records establish his presence in Arizona, his approximate age, his wife, and his children. But the census cannot tell us his parents' names, his mother's maiden surname, the village his family called home, or the community he came from. The baptismal record gives us all of that. Gerardo Vasquez, his father, Concepcion Ibarra, his mother, with her own surname preserved, and Tomas Carrillo, the godfather, which Claude correctly identified as not a minor detail. In Mexican Catholic tradition, the Campendrazo Network, the system of godparent relationships, typically connected families of similar social standing who often came from the same community. Tomas Carrillo may well have been Gerardo Vasquez's neighbor, his cousin, a man from the same village. If Tomas Carrillo appears in other records from the same parish, that gives us not just another research thread, but a window into the community Esteban came from. From before he was old enough to cross any border. On what to look for next, Claude recommended three specific directions. A marriage record for Gerardo Vasquez and Concepcion Ibarra in the same parish, which would extend the family line another generation and anchor the specific village. Any records for Tomas Carrill in the same parish collection, browsable in exactly the same way. And the Mexican border crossing records held at NARA, which document some Mexican nationals crossing into Arizona in the early 1900s, with birthplace information more specific than anything a census enumerator recorded. This is GPS element 3 in action, not treating any single record as the whole answer. Asking what all of them together can reasonably tell us and being clear about where the evidence is solid and where it still needs work. Let me tell you where we are now versus where we started. We started with one word, Mexico. We have arrived at Estevan Vasquez Ibarra, born 1866, baptized in a parish in the Sonoran Highlands, son of Gerardo Vasquez in Concepcion Ibarra, with a godfather named Tomas Carrillo who opens a research thread we haven't even fully followed yet. A family with roots in Sonora that the US Census never thought to ask about and the church never stopped recording. Think about what that baptismal register represents. A priest sat down in eighteen sixty six and wrote those names in a book because the church had been writing names in books since the fifteen hundreds. He was doing what his predecessors had done, what the instructions from the diocese required, what the sacramental tradition of the church demanded. He wrote Gerardo Vasquez because that was the child's father. He wrote Concepcionibara because that was the child's mother. And he used her surname, her family name, the name she had carried before she married into the Vasquez line. I want to sit with that for a moment because it matters. In the American documentary record, women disappear. A widow appears in the census as the head of household without a maiden name. A mother appears as a relationship, an age, a number in a household count. The administrative habits of the American state had very little use for who a woman had been before she became someone's wife or mother. Her identity before marriage was not the census' concern. The Catholic Church thought differently. The sacramental record required both parents because both parents were spiritually accountable for the soul being brought into the faith. Concepcion Ibarra is in that register because the church understood her as a person, not just as Gerardo Vasquez's wife. Concepcion Ibarra, a name the American census never wrote down. Now it's written down again. I want to say something now to the listeners for whom this episode is personal. For too many families with Hispanic or Latino heritage, there's been a quiet message across generations that the records on the Mexican side are somehow harder or less accessible or require resources or skills that most people don't have. That the ancestors on the other side of the border are less findable than the ones who stayed north of it. That is not true. The Catholic Church kept extraordinary records in colonial Mexico. Family Search has been digitizing those records for decades. And with AI handling the transcription and the correlation work, the barriers of language and handwriting are not what they once were. Your ancestors are in those books. They have been there the whole time. The next steps for this particular research are clear and specific. A marriage record for Gerardo Vasquez and Concepcion Ibarra in the same Sonoran parish would extend the family line another generation and anchor the specific village. The Compadrazgo thread through Tomas Carillo is worth following in the same browsable collection. And the NARA border crossing records for Arizona in the early 1900s might give us Estevan's exact birthplace from a form he filled out himself. But for the question that started today's work, where did Esteban Vasquez Ibarra come from and who were his parents? We have the answer. He came from Sonora. His parents were Gerardo and Concepcion. His family had roots in Mexico that predate civil registration, predate Arizona statehood, and predate the border where it currently stands. And when you ask which side his family was from, the answer is both. It always was. The border moved. The family didn't. Your homework. Go to familysearch.org and navigate to the Mexico collections. Find the relevant state for your family. Browse the Catholic records. If you find a record image you can't read, bring it into Gemini via AI Studio and use a version of the transcription prompt we walked through today. And if you get a transcription, bring it into Claude with everything else you know and ask it to help you understand what the records are actually telling you. The workflow is replicable. The records are waiting. For my Australian and UK listeners, the techniques we cover today translate directly to your Catholic record research. Australian researchers tracing Irish, Italian, or Southern European ancestry should look to the Catholic diocesan archives held in each state, which hold parish registers going back into the 1800s alongside Find My Pass Irish and British Catholic collections. For UK listeners, the National Library of Ireland's Free Parish Register database at registers.nli.ie holds digitized Catholic baptisms and marriages for most Irish parishes up to approximately 1880. The records are browsable by parish but not searchable by name, so knowing your ancestors' county and parish is key. Find my past Catholic Heritage Archive holds Catholic records for England and for all eight Scottish dioceses. For civil registration and Church of Scotland Records, Scotland's people, Scotland'speople.gov.uk is the primary source. The Gemini handwriting transcription workflow works just as well on 18th century Irish parish script as it does on Spanish language Sonoran Records. Same approach, different archives. Thank you so much for listening to Ancestors and Algorithms. If this episode opened a new direction in your research, please take a moment to leave a review wherever you listen to podcasts. And if you know someone who has been trying to get past that border wall in their family history, please share this episode with them. That is the best way to help our community grow. For my Patreon members, the companion guide for this episode is now in your library. It includes the full Spanish language transcription workflow, 12 advanced prompts for navigating the Mexican civil registration system, a guide to the Campan Drago's research strategy for extending family networks, and a GPS research checklist built specifically for border crossing genealogy. And if you've been thinking about joining, I host monthly live Q ⁇ A sessions on YouTube exclusively for members. Head to ancestorsandai dot com to learn more. For everything you need, including every episode, our private community, companion guides, and the research lab, head over to ancestorsandai.com. It's all right there waiting for you. I'm your host Brian, and I will see you next week for another journey into the past powered by the future. Until then, happy researching.