An "officer-involved shooting" means a police officer shot someone. That is the whole translation. The phrase exists to avoid saying it in that order, because in that order the sentence has a subject who did something, and the subject is a police officer. Swap it back and every version of the story reads differently: "Man dies in officer-involved shooting" becomes "Police officer shoots and kills man." Same facts. Same case file. Completely different sentence, and — as it turns out — a measurably different reader. The phrase is not a journalism invention. It is a police phrase that journalism adopted. In a Columbia Journalism Review piece on August 7, 2020, Mya Frazier traced the first known newspaper appearance to the summer of 1972, in the Independent Press-Telegram of Long Beach, California, quoting a Long Beach police detective. By 1974 the Los Angeles Police Department had institutionalized it to the point of having an official supervisor who carried the title. The lexicographer Ben Zimmer told CJR, "We can't know for sure when it was first used, but it probably spread from the LAPD." From there it moved the way police jargon always moves — out of the press release, into the notebook, onto the wire, into the headline. * Grammatically, the trick has a name, and it was invented for politicians, not cops. The political analyst William Schneider coined "the past exonerative tense" to describe constructions like "mistakes were made" — the Watergate-era non-apology Richard Nixon's press secretary Ron Ziegler made famous, and the formulation Ronald Reagan reached for over Iran-Contra. The grammar admits that a thing occurred while carefully declining to say who occurred it. In a Washington Post column on July 14, 2014, "The curious grammar of police shootings," Radley Balko took that same observation and pointed it at police statements, where the identical machinery gets used for identical reasons. There are really three separate moves inside the phrase, and it helps to see them apart. The first is the passive voice: "a suspect was shot," which lets you finish the sentence without naming a shooter. The second is nominalization — turning the verb "shot" into the noun "shooting," an event that simply happened, like weather. The third is the vaguest and most important word in the whole construction: involved. Involved how? Claire Moore-Cantwell, a linguist at UCLA, put it to CJR plainly: "You still don't know who is doing the shooting." An officer standing at the perimeter is involved. An officer who fired eleven rounds is involved. The phrase treats those as the same fact. * The obvious objection is that this is a fuss about wording. That objection is testable, and it has been tested. In an NBER working paper released in July 2022 — "Officer-Involved: The Media Language of Police Killings," by Jonathan Moreno-Medina, Aurelie Ouss, Patrick Bayer and Bocar A. Ba, later published in the Quarterly Journal of Economics — the authors examined US television news coverage and found that outlets use responsibility-obscuring structures (passive voice, nominalizations, intransitive verbs) more often when police kill someone than when a civilian does. Then they ran an experiment on readers. People shown the obfuscating version were less likely to hold the officer morally responsible, less likely to want penalties, and more likely to support police funding — with the largest effects in cases involving unarmed victims. The paper also found that obfuscating language went up, not down, in exactly the cases where clarity mattered most: when the person killed was unarmed, and when video of the incident existed. And when the authors looked for the source of the pattern, they did not find it in outlet ownership or political slant. They traced it to the police departments' own narratives — which is to say, to the press release the newsroom started from. That is the part worth sitting with. This is not a story about biased reporters. It is a story about what happens when a story's first draft is written by the institution the story is about, and nobody rewrites the verbs. * So how often does it actually happen? On January 13, 2022, Brandon Soderberg and Andy Friedman published an analysis (in HuffPost and via the Garrison Project) of roughly 136,000 newspaper articles from 35 major US outlets, filed under LexisNexis's "Death and Injuries By Police" category, covering 2000 to 2021. In 2000, "officer-involved" showed up in about 5 percent of those articles, averaging 22 uses a month. By 2010 it was 11 percent and 34 uses a month. The peak came in July 2016: 447 uses in a single month, in 13 percent of articles. In May 2020, the month George Floyd was killed, it appeared 130 times, in 8 percent of articles. And there was a correction, of sorts. On August 24, 2020, the AP Stylebook — the rulebook most American newsrooms actually run on — issued guidance: "Avoid the vague 'officer-involved' for shootings and other cases involving police. Be specific about what happened. If police use the term, ask: How was the officer or officers involved? Who did the shooting? If the information is not available or not provided, spell that out." AP Stylebook editor Paula Froke told the researchers the guidance means don't use it, with narrow exceptions when space is tight. In October 2022 AP extended the same advice to "police-involved." Usage did fall after the guidance. Then it climbed again in 2021. And the source kept supplying it: the LAPD's own December 2021 statement about a shooting in North Hollywood called the incident an "officer-involved shooting incident," which is the 1970s phrasing, intact, fifty years later. * None of this is a small vocabulary quarrel, because the underlying numbers are not small. Mapping Police Violence recorded 1,202 people killed by police in the United States in 2025 — more than three a day — down about 5 percent from 1,270 in 2024, the first year-over-year decline since 2021. Black Americans were killed at nearly three times the rate of white Americans relative to population. (The Washington Post's Fatal Force database, which tracked police shootings from 2015, stopped adding new cases on January 1, 2025, so the count now leans on nonprofit trackers rather than a major newsroom.) Every one of those is a sentence somebody has to write, and the choice of verb is made hundreds of times a year. Here is the useful part for you as a reader. When you see "officer-involved shooting," or "the weapon discharged," or "a struggle ensued," or "the man died following an encounter with police," try one substitution: put a subject in front of the verb. Who discharged it? Who struggled with whom? If the story can answer that, the writer chose not to. If the story can't answer it, that absence is itself news, and a good outlet will say so out loud — "police have not said who fired" is a real sentence that reporters are allowed to write. And notice which direction the vagueness runs. The same story that says a man "was shot in an officer-involved shooting" will usually manage a crisp active sentence for the man: he "lunged," he "fled," he "refused commands." When one side of an incident gets verbs and the other side gets weather, that is not neutrality. That is a choice about who has agency, and it was made before the reader arrived. * Wesley Lowery, who covered police killings for the Washington Post and now writes for the New York Times, has offered the entire fix in five words: "the police shot someone." It is shorter than the euphemism. It requires no extra sourcing. It commits the outlet to nothing it doesn't already believe happened. The reason it still loses is the reason most newsroom language loses: not conspiracy, but gravity. The press release arrives first, it arrives written, and a night-shift reporter working a story that must post in twenty minutes reaches for the words already on the page. Attribution verbs get chosen fast. Nobody in that chain intends to exonerate anyone. That is precisely why the AP had to write it down as a rule — because at deadline, the default is whatever the source handed you. If this looks familiar, it should. It is the same machinery as the "death tax" versus "estate tax" fight: identical facts, opposite gravity, all of it in the wording. It is why it matters that headlines are almost never written by the reporter who did the work, and why a story built from an institution's own press release deserves a second read. The euphemism is not hiding a fact from you. It is hiding a subject — and the subject was the story.