You Don't Get a Vote on Whether Someone Is Telling the Truth About Their Own Drinking

Katherine walked into her First Step meeting with an honest story. Her counselor read it first and told her it wasn’t honest enough. So Katherine added things — invented details, sharpened the rock bottom, made the story sound like the story a counselor expects to hear from someone with a drinking problem. The true version wasn’t believable. The embellished version was. Nobody asked her to lie. The system just wasn’t built to hear the truth if it came in the wrong shape.

Three thousand miles and seven years away, Brad Pitt told a reporter he’d gone back to wine. “I was sober for seven years,” Pitt said in an Esquire interview reported by The Hollywood Reporter this month. “And then I got back off the wagon.” He drinks “in a more restrained manner” now — a few glasses, “professional about it.” The internet didn’t receive this as news about one man’s life. It took it as a referendum on whether anyone with a history of heavy drinking can be trusted to describe their own drinking again — the exact question at the center of a sharp new essay by philosophy professor Peg O’Connor.

When you’ve had a drinking problem, other people don’t just doubt your future — they retroactively doubt your past, your present, and your ability to know either one. That’s not a side effect of stigma. It is the stigma.

The credibility gap has a name, and it isn’t yours to fix alone

O’Connor, who teaches at Gustavus Adolphus College and writes the Philosophy Stirred, Not Shaken column, uses Pitt’s disclosure as her case study, and pairs it with a harder one: Matthew Perry, who wrote in his memoir that he’d been sober for eighteen months before he died from the acute effects of ketamine, with investigators later describing an addiction he’d been concealing even from people close to him. In her piece, O’Connor quotes clinical social worker Joe Schrank putting the skeptic’s case as bluntly as it gets: “Typically speaking, alcoholics are just a horrible source of information about themselves.” Schrank points to Perry as the proof. Someone said he was sober. He wasn’t. Case closed, or so the logic goes.

Here’s the term worth learning, because you’ll use it the next time someone waves away what you say about your own life: credibility isn’t a fact about the speaker, according to O’Connor — it’s a relationship. Being a reliable narrator of your own experience doesn’t mean much if nobody around you is willing to take the claim seriously in the first place. A person can be telling the exact truth and still lose, every time, if the room has already decided people like them can’t be trusted to know themselves. That’s the credibility gap: not that people with addiction histories lie more than anyone else, but that they’re assumed to be lying — or self-deceived, which O’Connor treats as a separate and gentler failure than dishonesty — before they’ve said a word.

The awkward part, which O’Connor doesn’t dodge, is that Schrank isn’t inventing his skepticism from nowhere. He’s the one in the piece who actually cites the World Health Organization’s own position that moderate drinking after extended abstinence is a real, viable outcome for some people — and then argues that self-report is still too shaky to know who “some people” actually are. That’s the tension the whole essay sits inside: an international health body says moderation counts as recovery for a subset of people, while the fellowship most Americans still associate with recovery, Alcoholics Anonymous, has built sixty-plus years of practice on the premise that self-report from someone with a drinking history can’t be trusted to draw its own line. Both institutions can’t be fully right. Somebody’s credibility model is wrong, and it’s usually applied to the person with the least power in the room.

A tale of two disclosures — and why only one got the benefit of the doubt

Put Pitt and Perry side by side and the credibility gap gets easier to see, because the two men made almost the same claim and got opposite receptions before either outcome was even settled. Pitt said, in effect, “I’ve recalibrated my relationship with alcohol and I’m managing it.” Perry said, before his death, “I’ve been sober for eighteen months.” Perry turned out to be wrong — tragically, by his own account and the investigation that followed. Pitt might turn out to be wrong too; nobody knows yet, including Pitt. But notice what happened to the believability of each claim. Perry’s turned out false and got treated, after the fact, as proof that people with his history can’t be trusted to report on themselves at all — a lesson applied backward onto everyone, not just him. Pitt’s hasn’t been tested by time yet, and it’s already being treated with suspicion in advance, on the theory that a person seven years sober forfeits the right to describe his own drinking without a professional’s sign-off.

Pitt might turn out to be wrong too; nobody knows yet, including Pitt.

This is what a relational model of credibility predicts. It isn’t really about facts on the ground. It’s about who the audience already trusted walking in. A person with no addiction history who says “I’m going to have a couple of glasses of wine and see how it goes” gets taken at face value. A person with Pitt’s history saying the identical sentence gets a think piece. That gap is the whole argument, and it doesn’t close just because one case eventually turns out fine and the other doesn’t — because the doubt gets applied before anyone knows which is which. If you’ve ever had a doctor, a family member, or an old friend flinch at your version of your own night, your own week, your own decision to have one drink at a wedding — you already knew this before O’Connor put a name on it. The gap isn’t a hypothesis to you. It’s Tuesday. It’s also the reason this outlet runs a whole lived-experience section instead of leaving that story to clinicians alone.

The historical throughline: this fight is thirty years old and it has a body count

This isn’t a new argument — it’s the same fight Audrey Kishline started in 1994, when she founded the first public group arguing that some people with drinking problems could learn to drink in controlled, limited amounts instead of quitting for life. She called it Moderation Management. Kishline had genuinely helped people, and abstinence-only critics called moderation a dangerous fantasy. In January 2000, Kishline concluded she’d been wrong about her own case, switched to abstinence — and two months later drove drunk into an oncoming truck, killing a father and his 12-year-old daughter. Her blood alcohol was more than three times the legal limit. The tragedy became, instantly, “proof” that the entire moderation idea was a lie — even though it was a story about one person’s relapse, not a controlled study of a method.

That history is why the current fight feels loaded before anyone says a word. Kishline’s crash didn’t just end her own life’s project; it got weaponized for twenty-five years as evidence that anyone claiming they might be the exception is lying to themselves. The WHO’s more recent stance that moderation can be a legitimate outcome after abstinence is, in effect, catching up to what Kishline argued three decades before her worst day — without ever fully reckoning with why her worst day happened at all.

The receipts say the doubters have it backward

If the working assumption is that people can’t be trusted to accurately describe their own drinking, the country’s actual drinking data makes that assumption look worse every year it’s tested. Gallup’s newest poll, published August 20 by Lydia Saad, found only 54% of U.S. adults now drink at all — tying last year for the lowest share Gallup has recorded since it started asking in 1939, down from 62% in 2023. Among people who do drink, just 13% say they sometimes overdo it — also a record low. Fifty-one percent now believe even one or two drinks a day is bad for their health, and 17% say they’ve swapped in a nonalcoholic drink sometime in the past year. Millions of self-reports, moving in the same direction, year over year — that’s not a population of unreliable narrators. That’s a population self-correcting faster than the institutions built to distrust them. For more of what that recalibration looks like up close, see our ongoing alcohol coverage.

The clinical data backs it up. A JAMA Psychiatry study published August 19 by lead researcher Beth Han, drawing on more than 186,000 U.S. adults surveyed between 2021 and 2024, found alcohol use disorder — what clinicians call it when drinking has crossed into causing real harm to your health, relationships, or obligations — actually fell in women, from 9.7% to 8.2%, while holding roughly flat in men overall. The one group where it climbed sharply was young men 18 to 20, up from 6.9% to 10.4%, a genuine warning sign worth its own conversation elsewhere. But the top-line story in Han’s data is a national decline in alcohol use disorder at the same moment Gallup is recording record-low drinking and record-low overdrinking, reported by the very people being told they can’t be trusted to report it.

The one group where it climbed sharply was young men 18 to 20, up from 6.9% to 10.4%, a genuine warning sign worth its own conversation elsewhere.

O’Connor’s essay ends on the philosophy. This one should end on the arithmetic. A country full of people self-reporting less drinking, less overdrinking, and more discomfort with the whole enterprise is not a country of unreliable narrators — it’s a country full of people doing exactly what Pitt described and what Kishline attempted and what Katherine tried to say honestly before her counselor told her to make it sound worse: paying close attention to their own experience and adjusting. The credibility gap was never really about whether people with a drinking history can be trusted. It was about whether the rest of us are willing to find out.