Almost every serious my family tree project fails the same way — the researcher spends weeks on archives, DNA tests, and online databases before ever sitting down with the elders who actually remember the family. Then the elders die and the memories die with them. Every year of delay costs a generation of family history.
This guide walks the specific discipline of building a family tree from living memory first: how to plan the interviews, what to ask, how to record and transcribe them, and how to convert the raw memory into a working tree node before layering archive records on top. Written for Tamil families where oral history often carries pre-1900 detail that no archive record preserves.
Why living memory comes first
Three practical reasons to build the initial tree from oral history, not records:
- Perishability. Elders in their 80s and 90s carry family history that dies with them. Records survive; memory doesn’t. The record can wait a decade; the interview can’t.
- Density. One 90-minute conversation with a grandmother often yields more genuine family history than 40 hours of database search. She remembers her grandmother’s cousin’s marriage; no record captures that specific fact.
- Direction. Oral history tells you WHERE to search records. Without direction, archive search is guessing. With the village name, community, and rough era from an elder’s memory, the archive becomes tractable.
Who to interview first
Rank interviews by both age and connection depth. Priority ranking for building my family tree from living memory:
- Priority 1: the oldest living relative on both maternal and paternal sides. Age is the biggest factor — every year risks losing them.
- Priority 2: siblings of your grandparents (great-aunts / great-uncles). They typically carry the same generational memory as your grandparents but often with different specific details.
- Priority 3: your parents’ cousins from both sides. Extended family branches carry information the direct line doesn’t.
- Priority 4: your own siblings, cousins, and older family friends. Cross-check details; some memories differ across cousins.
The absolute worst mistake is starting with Priority 4 and never getting to Priority 1. Book Priority 1 interviews now, even if you’re not ready — they can wait for you to prepare questions; they can’t wait for you to feel expert.
How to structure the interview
A good family tree oral history interview runs 60-120 minutes. Split into three phases:
Phase 1: Warm-up (15 min)
Start with easy, non-genealogical questions. What was your childhood home like? What did you eat as a child? What songs did your mother sing? These questions relax the elder and often surface incidental family detail (mother’s specific temple, father’s specific occupation, village landmarks) without the pressure of “give me names and dates.”
Phase 2: Systematic naming (45-60 min)
Move to structured naming questions:
- Your parents — full names, birth villages, occupations, when they died, where they’re buried
- Your grandparents (both sides) — same detail
- Your great-grandparents — as much as you remember
- Your parents’ siblings — names, spouses, where they moved to
- Your grandparents’ siblings — same
- The family’s original village (if diaspora) — specific village name, taluk, district, when the family emigrated
- Any family stories about migration — which ship, which port, which year, which destination
Ask open questions (“Tell me about your grandmother”) rather than yes/no questions. Follow up on details (“You mentioned Wandiwash — do you know what taluk that’s in?”). Record every name mentioned even if you don’t know how it fits yet.
Phase 3: Artifacts (15-30 min)
Ask about physical family artifacts:
- Old photographs (with permission to scan)
- Family Bible or kula deivam register
- Land documents, patta records
- Funeral cards from past deaths
- Old letters, especially from India-to-diaspora correspondence
- Marriage invitation cards from past weddings
- Passport, ration card, voter ID, service certificates
Ask if you can borrow or copy the artifacts. Scan everything at high resolution — the elder holding the artifact today may not be here next year, and the artifact may not survive whoever inherits it.
Recording setup that actually works
Bad recording ruins good interviews. The minimum reliable setup:
- Audio recorder: a smartphone with a voice-recorder app works fine (native iOS Voice Memos or Android Recorder). Test the recording quality in a room before the interview.
- Backup recorder: a second device recording in parallel. If the primary fails, you don’t lose the interview.
- Video (optional): captures body language, gestures, facial expressions that carry meaning. Useful for the elder to point at photographs or family artifacts during the interview.
- Written notes: keep them minimal during the interview (don’t break eye contact) but do note key names, dates, places as they come up. Names spelled out are especially valuable for Tamil-script transliteration checks later.
- Permission: ask explicitly whether recording is OK. Some elders are self-conscious; some prefer only audio. Respect their preference.
Transcribing and extracting facts
Within 24-48 hours of the interview:
- Back up the recording in two locations (cloud + local).
- Transcribe verbatim, or at least generate a rough timestamped outline. Free auto-transcription tools (Otter, Descript, Whisper) work adequately for English + are decent for Tamil-language content.
- Extract every named person into a spreadsheet: Name / Relationship to interviewee / Village / Date range / Sources for this claim (with the interview being the source).
- Extract every named place: village names, districts, ships, plantations, destination countries.
- Flag contradictions against previously-known family history. Contradictions are research leads.
Converting oral memory to tree nodes
Not every remembered person becomes a tree node immediately. The rule:
- Add as verified node: person named + relationship confirmed + one supporting source (another elder, a document).
- Add as unverified node: person named by one elder only. Note “source: interview with [name] on [date]”.
- Add as lead only: vague mention (grandfather had a sister but her name is forgotten). Note it, but don’t create the tree node until a specific name surfaces.
An oral history family tree built this way is honest about what it knows. Future descendants can see which ancestors are verified vs. remembered vs. speculated.
Cross-referencing memory against records
Once oral history is captured, the record trail becomes tractable. For each named village, search:
- Tamil Nadu voter rolls (Electoral Commission online portal)
- Land records at patta.tn.gov.in (current holdings — may reveal descendants of the extended family)
- District Registrar of Births and Deaths (for post-1890 civil registration)
- The specific temple mentioned (kula deivam records, festival contribution registers)
Discrepancies between oral history and records are common and usually resolvable:
- Elder says “grandfather was 40 when he died” but the death certificate says 47 — memory error, weight the certificate.
- Elder says “our village was Chetput” but records only show “Chennai” — Chetput is a specific neighbourhood of Chennai; both are right.
- Elder says “grandfather emigrated in 1885” but arrival record shows 1887 — memory error on the year; weight the record.
The goal isn’t oral-history-vs-records winner-take-all. It’s using each to correct and extend the other.
Return visits
First interviews rarely capture everything. Schedule a return visit within a few weeks with:
- A specific list of follow-up questions (based on gaps in the first transcript)
- Any records / photographs you’ve found since that might prompt further memory
- Photographs of the extended family from other branches (elders often recognise relatives they hadn’t mentioned)
Two or three interviews with the same elder over a few months typically produces 2-3x the content of a single long session. Memory is layered — some details surface only when triggered by a specific question or an image.
Building my family tree from living memory is not just a step in the process — it’s the foundation. Every subsequent archive search is more productive because you know what to look for. Every DNA test is more meaningful because you have context to interpret the results. Every future descendant inheriting the tree has a living link to ancestors who would otherwise be just names on paper.
