There is a sentence that appears in British political journalism almost every week, and it tells you almost nothing.

"Sources close to the chancellor said..." It sounds authoritative. It works as fog. You cannot weigh the claim, because you do not know who made it, how they know, or what they want. The confident tone is doing the work that evidence should be doing.

Attribution is not a formatting detail. It is what allows a reader to check a claim, and what allows the next reporter to build on it. When it goes soft, everything downstream goes soft with it. This is an argument for sharper, more honest attribution — not for scrapping anonymity.

The same four words, a dozen different meanings

Consider what sits underneath "a source close to the leadership". It might be a special adviser who was in the room when the decision was taken. It might be a minister who heard about it afterwards, and is unhappy. It might be a donor who has the leader's ear on a Friday night and a grudge by Monday morning. It might be someone repeating what they read in a group chat.

All four could be stories. None of them carries the same weight. The label flattens that difference into a phrase that reads like proximity to power and could mean almost anything. Readers have no way to tell a first-hand account from a fourth-hand one, and after a while they stop trying.

Anonymity is not the problem

There are good reasons to protect a source. A civil servant who has seen a document and would lose their job for describing it. A junior aide who witnessed bullying. A whistleblower inside an organisation whose first instinct is to find the leaker. A reporter in a country where naming a contact puts them in danger. In those cases anonymity is not a shortcut — it is the only way the story reaches the public.

The failure is not that anonymous sources exist. It is that so many of them arrive with no context at all, and that newsrooms apply the same thin phrase to a frightened official and to a spin doctor settling a score.

Why vagueness survives

Part of it is competition. Push a source for a better description — "a Treasury official who has seen the costing" — and there is a real chance they take the story to someone who will settle for less. Under deadline, and against a rival's splash, the weaker label is easier. It is also safer: a vague source cannot be checked by anyone outside the conversation, including the desk.

Part of it is that vagueness suits the people doing the briefing. Anonymous quotes are how leadership challenges are tested, how a minister is undermined by their own colleagues, and how a tax rise is floated to see whether it explodes. The briefer gets the effect without the fingerprints.

Then there is the recycling. A single briefing becomes a story, the story becomes "reports suggest", and the reports become "multiple sources" in tomorrow's copy. A claim can travel a long way from the person who first said it. That is not a reporting technique; it is how a thin claim acquires a thick reputation.

The bill arrives later

Every vague attribution spends a little of the same credit. Readers notice when a dramatic claim built on unnamed sources quietly evaporates, and they remember the splash rather than the clarification. The original gets the front page; the correction gets a paragraph at the bottom of page twelve.

It also makes life harder for the sources who genuinely need protection. When an audience has learned to discount anonymous quotes, the whistleblower with documents and a mortgage walks into a more cynical room than they deserve.

What clearer attribution would look like

None of this requires identifying anyone. It requires describing them better. Four habits would change a great deal:

• Describe proximity, not just existence. "A Home Office official involved in the discussions" tells the reader where the information comes from. "A source close to the Home Office" tells them nothing they can use.
• Describe what the source stands to gain. A supporter of one leadership campaign is not a worse source than a neutral one, but the reader deserves to know which they are getting. Motive is information, not a disqualification.
• Say why anonymity was granted. One line does it: they are not authorised to speak publicly, they fear losing their job, they would face disciplinary action. If none of those applies, ask why the source is hidden at all.
• Do not multiply one briefing into many sources. If every version of the claim traces back to the same conversation, say so. "Two people familiar with the briefing" is honest. "Sources", implying a chorus, is not.

If a source cannot be described in any of those ways, that is worth noticing. The story may still be publishable, but it probably belongs lower down — or in a piece that flags the claim as unverified.

Five questions to ask when you see the phrase

Readers can do some of this work themselves. Editors should be doing it before publication.

1. Is this person in the room, in the building, or passing on a rumour?
2. What do they gain if the claim is believed?
3. Does the piece explain why they are anonymous?
4. Is anything independently confirming it, or is every version traceable to one briefing?
5. If it turns out to be wrong, will the correction get the same prominence as the original?