What to Do When You Hit a Brick Wall in Genealogy
Break a genealogy brick wall into a testable identity question, audit the evidence, expand the person’s network, and search records that can distinguish candidates.
A genealogy brick wall is usually not “no records exist.” More often the question is too broad, a hidden assumption is wrong, the person changed name or place, two identities were merged, or the decisive record lies in another jurisdiction or record type. Stop repeating surname searches and rebuild the problem from evidence.
Write one precise question
Replace “Find Patrick Kelly’s parents” with:
Who were the parents of Patrick Kelly, a dock labourer who married Anne Cain in Liverpool in 1887 and appeared at 42 Regent Road in the 1891 census?
Include identifiers that distinguish him from namesakes. Then list what a successful answer must explain: age pattern, occupation, spouse, children, addresses, birthplace, religion, witnesses, and associates.
Audit every fact
Create a timeline and cite the actual records. For each statement, note whether you viewed an original, image, transcript, index, family tree, or oral account. Reopen images and read headings, marginal notes, adjacent entries, and the whole household.
Check for facts copied from one another. Ten public trees with the same birthplace may trace back to one unsourced submission. Remove unsupported claims temporarily and see whether the research problem changes.
Challenge the identity
Ask whether all records concern one person. Build separate timelines for same-name candidates. Look for impossible overlaps, abrupt occupation changes, inconsistent relatives, and different signatures.
Also challenge the starting event. A death age may be wrong; a marriage may be second rather than first; a “birthplace” may mean parish, nearest town, or nationality; a stepfather may be named as father.
Search the person’s network
Research siblings, spouses, children, witnesses, godparents, neighbours, employers, travelling companions, and informants. People migrated and appeared in records as groups. A sibling’s obituary, marriage, military file, or probate record may name the parents that the target’s records omit.
Create a network table with names, relationship or association, dates, places, and sources. Repeated witnesses or neighbouring households often reveal kinship.
Broaden names methodically
Search spelling variants, phonetic equivalents, initials, aliases, married names, patronymics, translations, and common transcription errors. Browse images when indexes fail. Search by spouse, child, occupation, address, or birthplace with the surname omitted.
Do not broaden everything at once. Record each search range and filter so a negative result has meaning.
Follow jurisdiction and record creation
Determine which authority created the record at that date. Boundary changes may move a parish, court, registration district, or county archive while the family stayed put. Use period maps and gazetteers, then consult the catalogue for the responsible repository.
Ask whether records survive and whether they are online. Digitised collections are only a fraction of archival holdings; some indexes omit restricted years or entire record series.
Build a record-type matrix
List records that could answer the question and their status:
| Record | Expected clue | Coverage/location | Status |
|---|---|---|---|
| Marriage register | father, witnesses, residence | parish archive | image needed |
| Civil certificate | age, father, occupation | registry | order copy |
| Census | birthplace, household | national archive | searched 1891 only |
| Probate | relatives and property | court archive | catalogue not checked |
| Newspaper | obituary, migration, court | local titles | surname search incomplete |
Prioritise records likely to discriminate between hypotheses, not simply add another approximate age.
Use negative evidence carefully
Absence matters only when the source should contain the person, coverage is complete, and the search was thorough. “Not found online” is not proof. Record collection limits, variants, browse ranges, missing pages, and access restrictions.
A household missing from one census might be abroad, mistranscribed, on a damaged page, or under another name. Several independent absences can support a conclusion when expectations are well defined.
Form competing hypotheses
Write at least two explanations:
- Patrick was born in County Mayo to Michael Kelly.
- Patrick was born on the Isle of Man to John Kelly and later reported Ireland loosely.
- The Liverpool husband is not the same Patrick as the earlier record.
For each, list supporting evidence, conflicting evidence, and a record that could falsify it. Actively searching for disproof reduces confirmation bias.
Know when DNA may help
DNA can support biological relationship questions when tested with consent and analysed alongside documented trees. It rarely names an ancestor by itself. Work with shared matches, relationship ranges, endogamy, pedigree collapse, and multiple hypotheses. Protect living people and be prepared for unexpected parentage.
Ask for a fresh review
Give another researcher the precise question, timeline, evidence table, search log, hypotheses, and source images. Do not give only your preferred answer. A fresh reader may spot an assumed relationship, jurisdiction error, or unread marginal note.
Define progress realistically
A brick wall is moving when you have corrected an identity, excluded candidates, identified missing coverage, found a new network, or narrowed the next record. Write a short proof summary or status report even if the answer remains unknown.
The goal is not endless searching. It is a transparent, repeatable investigation in which every next step is chosen because it can change the conclusion.