The Case Closed. The Mug Shot Still Wins the Search.

In 2010, Connecticut police arrested Lorraine Martin on drug charges. Prosecutors later declined to pursue the case, and the state’s erasure law treated her arrest records as erased. News reports about the arrest remained online.

Martin sued several publishers, arguing that the reports had become false after the state erased the record. In 2015, the federal appeals court rejected that argument. The reports had accurately described an arrest when it occurred, the court held, and Connecticut’s legal erasure did not rewrite the historical event.

The resulting appellate opinion, a public PDF carrying Martin’s full name, explains the contradiction with unusual clarity. The state can declare an arrest erased inside its own legal system while the internet continues to treat the accusation as durable biographical information. One institution changes the record. Another preserves the story about it.

That PDF is the anchor here because it shows the problem without requiring a shadowy data broker or an obvious technical failure. The system can operate as designed and still produce a permanent accusation. Once arrest information leaves the government portal, every later correction must chase copies that have different owners, incentives, formats, and legal protections.

A mug shot makes the same journey with more force. It looks official because it is official. It also travels without the procedural context that might follow: charges dismissed, identity mistaken, evidence excluded, conviction reversed. Search does not place those events in order.

It places pages in rank.

The accusation has a distribution system

An arrest can begin as a row in a sheriff’s booking log, an entry on a jail roster, a police release, or a photograph attached to a county page. A court case generates another trail: docket entries, hearing calendars, complaints, motions, opinions, and sometimes transcripts. Public access supports scrutiny of police and courts. It also supplies clean material for businesses built around names.

Commercial sites can collect that material through downloads, public feeds, or scraping, which means software automatically copying information from web pages. The copied record then becomes a new page with the person’s name in its title, the alleged offense in visible text, and a stable address that search engines can index.

The economics favor intake. A site can publish large volumes of arrest pages from standardized government data, place advertising around them, route visitors toward paid background products, or use the pages to capture search traffic. Updating each record after prosecutors amend charges or a judge seals the file produces less traffic and demands more work. The accusation arrives in bulk.

The correction arrives as paperwork.

Some notorious mug-shot businesses added a harsher incentive by charging people to remove booking photographs. Public reporting by The New York Times in 2013 documented the pay-to-remove market and the pressure it placed on search engines and payment processors. California authorities later accused the operators of Mugshots.com of running an extortionate removal scheme.

Laws in several states now restrict charging for deletion or require certain sites to remove records under defined conditions.

Those interventions addressed the ugliest transaction. They did not eliminate the underlying supply chain. A page does not need to demand money from its subject to profit from an arrest, and a copied allegation can remain useful to its publisher as long as somebody searches the name.

Search ranks pages, not case histories

Search ranking is the system that orders results according to signals such as relevance, page structure, links, location, and perceived authority. It is not a docket clerk. When somebody types a full name, a page built around that exact name has a strong claim to relevance, particularly when the name appears beside official language and a government-produced photograph.

The final disposition may live somewhere else. A dismissal can appear as a short docket code. An expungement order may remove the source page altogether. A later opinion can sit inside a PDF whose title emphasizes the court rather than the person.

Records from prosecutors, trial courts, corrections departments, and police databases do not reliably share one public identifier that commercial publishers must use to reconcile the story.

This creates an asymmetry. The arrest page says what happened in ordinary language and carries a face. The correction often speaks in procedural shorthand, appears months later, and has no image. Search can therefore return a vivid allegation above the document that limits or defeats it, even though the state’s current file is more favorable to the person named.

Google has adjusted its treatment of mug-shot sites and offers a removal route for pages connected to exploitative deletion practices. It also removes some results when valid legal orders or privacy policies apply. De-indexing, which means excluding a page from search results without deleting it from the web, can reduce exposure.

None of that creates a general duty to pair an arrest with its outcome. A page can violate no stated search policy, make no removal demand, and remain factually accurate in the narrow Martin sense. The person then confronts a result that is old but not necessarily false, damaging but not necessarily prohibited, public but stripped of the process that followed.

The ranking system has no hearing where context prevails.

Erasure stops at the state’s edge

Expungement, sealing, and record erasure differ by jurisdiction, but each controls specified government records rather than every copy already released. A clerk may seal the docket. A police department may remove the booking page. A commercial publisher that copied the material earlier can claim a separate right to retain an accurate report about a public event.

The Martin opinion sits directly on that border. Connecticut changed the legal status of Martin’s arrest records. The Second Circuit did not require publishers to pretend the arrest never occurred. The judgment makes sense within defamation law, which asks whether a statement was false, yet it leaves the search subject carrying a state-created accusation after the state has withdrawn from the case.

There are legitimate reasons to preserve court reporting. Police misconduct, selective enforcement, prosecutorial abuse, and the treatment of public figures cannot be investigated through a record system that silently destroys history. A rule requiring every publisher to erase every report after an expungement would give the state considerable control over the archive of its own conduct.

That concern does not justify designing every booking log as a permanent, name-searchable feed. Public access and mass distribution are separate policy choices. A courthouse terminal, a downloadable dataset, an indexed page, and a licensed background-check product all expose the same record at different scales. Institutions often collapse those differences into the word public, then leave the person named to manage the consequences.

Historically, practical obscurity supplied friction. A record could be open to inspection without appearing instantly to an employer, landlord, date, customer, or neighbor. Digitization removed the trip to the courthouse and the need to know which court held the file. Search removed the need to know that a court file existed.

The correction needs its own infrastructure

The narrow repair is not a promise that search engines will understand justice. It is a chain of obligations attached to the data.

Government agencies could limit the indexing of booking pages, publish mug shots only where a defined public-interest reason exists, and distribute stable case identifiers with every released record. When a disposition changes, the same feed could carry the update in a machine-readable form, meaning structured data software can match to the original entry without interpreting a scanned order.

Contracts for bulk court and jail data could require recipients to ingest corrections, display dispositions beside allegations, and delete records after sealing where the law demands it. Search engines could downrank name-query pages that omit a documented dismissal while giving the linked disposition comparable prominence. Publishers would still retain archives. They would lose the privilege of presenting the first procedural event as the complete one.

Each measure has a cost. Courts would need compatible records, reliable identifiers, and staff able to correct mismatches. Small newsrooms could not be treated like industrial background-data vendors. Search companies would have to make judgments they prefer to describe as neutral ranking decisions.

The work is administrative, which is why institutions pass it downward.

The individual pays instead. They locate copies, submit identification, collect court orders, complete publisher forms, request search removals, and repeat the exercise when another domain republishes the same image. Reputation-management firms sell help with that labor, creating a second market around the state’s original disclosure.

Martin’s 2015 opinion remains useful because it records the point where the paper trail splits. One branch says the case ended and the record was erased. The other says the arrest happened and the report remains true. Search does not reconcile them.

It gives the better-optimized branch the first line of the biography.

Questions people ask

Why does a mug shot remain online after charges are dismissed?

The police or jail can remove its original page without controlling copies already taken by publishers, archives, and data brokers. A dismissal changes the case, but it does not automatically impose a deletion or correction duty on every private site holding the earlier booking record.

Does expungement remove an arrest from Google?

Not automatically. Expungement governs records covered by the relevant court order and state law, while Google indexes pages held by many independent publishers. A result may disappear after the source deletes it, a legal request succeeds, or Google applies a policy, but expungement alone does not synchronize those systems.

Are mug-shot websites allowed to charge for removal?

The answer depends on state law and the site’s conduct. Several states restrict mug-shot removal fees, and authorities have pursued businesses accused of using publication and payment demands as an extortion scheme. A ban on fees still may not require every publisher to delete an accurate historical record.

What would keep closed cases from outranking their outcomes?

Agencies would need to release stable identifiers and timely dispositions, commercial recipients would need enforceable update duties, and search engines would need to treat an omitted dismissal as a ranking defect. Without that chain, the booking photograph remains a complete web page while the correction remains a code in another database.