Newspapers do not forget: the pre-internet record in modern investigations

I spent the first part of my career in newsrooms, and the lesson I brought to Aldervane is that local papers are an index of ordinary life. They record who was born to whom, who married, who graduated, who was charged and which business opened on which street. Most of that record never made it into a search engine.

Three first names

Earlier this year, counsel for the liquidators of a collapsed investment fund needed to find a former employee whose evidence could decide a fraud claim. He had left the fund six years earlier, changed his surname and deleted his online accounts. Under the name we knew, his trail stopped completely.

The fund's old personnel file gave us his first name, his date of birth, a hometown and his mother's first name. Searching a digitized regional newspaper archive for those first names, in that region, within a year of his birth, returned four birth announcements. One matched the date. It named his mother, his father and the child, and the father's surname was one the fund had never seen. Under that surname, the witness appeared in a professional register and as the director of a small bookkeeping business. Counsel served a witness summons nine days after we were instructed. The full case is told in Three first names.

What the printed record holds

• Family notices. Birth, marriage and death announcements link people across generations and across name changes, and often name siblings, parents and hometowns.
• Community lists. Graduation lists, sports results, awards and club news place a person in a school, a town and a peer group at a known date.
• Local court reporting. Regional papers covered magistrates' and county courts in detail for decades. Charges and convictions from before the mid-2000s are often recorded nowhere else that is easy to find.
• Business news. Openings, closures, planning applications, licensing hearings and disputes with customers or councils.
• Public notices. Insolvency, probate, name change and licensing notices, published in newspapers and official gazettes because the law required it.

Why search engines miss it

Much of the archive is behind subscription paywalls that search engines cannot crawl. Where pages have been digitized, the text is produced by optical character recognition, which mangles names and is often not indexed at all. And a great many small papers have never been digitized. They exist on microfilm in a library or on bound volumes in the newspaper's own office.

That gap cuts both ways. It is why a contractor's 2014 fraud charge can be invisible to a standard screen, as it was in The lawsuit nobody read. It is also why people who believe they have erased their past have usually erased only the part that search engines could see.

How we search archives

1. Build variants. Names are listed with spelling variants, initials, maiden and married names and the errors that optical character recognition typically introduces.
2. Search narrowly. Combinations of first names, places and date windows often work better than a full name, especially for family notices.
3. Use several archives. Commercial archives, national and regional library collections and newspapers' own archives overlap only partly.
4. Go to the shelf. When the record exists only on microfilm or paper, a researcher visits or requests copies from the library or the paper's archivist. We request access openly, as researchers. We never misrepresent who we are or why we are asking.
5. Capture properly. Archive pages are captured with their source, edition and page, and hashed like any other evidence.

Court records: a company's past, in its opponents' words

Court records are the second half of the analog record and among the richest sources we use. Dockets list parties and dates. Filings describe disputes in detail. Exhibits attach contracts, invoices, emails and certificates that were never meant to be published. A civil claim by a subcontractor can name a company's site managers, describe its practices and reveal its relationships with other entities.

Access rules vary widely by jurisdiction and by court. We use records that are publicly accessible or that counsel obtains through proper process. Sealed, restricted and juvenile records are not touched. Many civil claims settle quietly, so the record of the claim may be the only public trace of the dispute.

Cautions

• A name is not a person. An archive match is a lead until dates, places and relationships confirm it. The four announcements in the case above shared first names. Only one shared a date of birth.
• A charge is not a conviction. Reports of arrests and charges must be followed to their outcome before they are reported as anything more.
• Old is not always relevant. Many jurisdictions treat old convictions as spent for most purposes. We report historical matters only where they are lawful to consider and relevant to the question, and we say how old they are.
• Proportion still applies. Family notices describe people who are not subjects. We take what the question requires and leave the rest.

Archive checklist