There's a specific kind of betrayal that only happens with actors. You watch somebody for years, you're certain they're English, and then you find an interview and the accent is pure Ohio.
It happens to me constantly. I'll have a whole mental file on someone and then a press junket torches it.
Daniel's been chewing on the same thing. His note is about how unreasonably good professional actors are at this, and the fact that behind the ones who take it seriously there's a person whose whole job is the accent. He wants the enterprise looked at properly. The trainers themselves, whether they're connected to the people who actually study speech, and then the big one underneath all of it: can anyone learn an accent, or is there a gifted subset and the rest of us are just standing there?
And whether it's trainable, innate, or the thing he suspects, which is both.
Right. Craft or lottery. Let's find out which.
Start with the people who do this for a living, because the method is more systematic than most people assume. A coach doesn't hand an actor a recording and say good luck. The first move is decomposition. You take a target accent and break it into constituent parts, the individual vowel and consonant sounds, the overall oral posture, the intonation.
Oral posture.
Where the whole apparatus sits at rest. Where the tongue lives, how much the jaw hangs, whether the sound is forward in the mouth or back. It's the thing you never notice about your own accent precisely because it's the default.
So the actor's copying a sound, and the coach is really teaching a posture.
And the reason you need a coach at all is that your own ears lie to you. You hear the target through the filter of your native categories. There's a line from a dialect coach about exactly this, that most people's ears and brains trick them because of interference from the mental categories their first accent established. You literally cannot hear the difference you're getting wrong.
The instrument's fault, not the player's.
William Conacher's method is the clean example. He spent eight years at RADA, he worked on The Crown, The Reader, Billy Elliot. First thing he does is get the right sample. And the right sample means matched. Same gender as the character, same period, same class position, because a received pronunciation from 1955 is a different animal from one from 1995.
He's auditioning the recording like it's the actor.
Then he obsesses over it. Sits with it long enough to notice things, writes all of it down, and then the actual craft step, which is finding the pattern. He's looking for half a dozen salient points. Six things he can hand the actor. His words are that he doesn't want to bombard people with changes, just the ones that give them a way in.
Six things for an entire accent.
That's the insight. The accent isn't six things, but the actor can't hold forty. Six is the number that gives them a foothold, and once they're standing on something, the ear starts doing its own work on the rest.
Which is not what a person expects. You'd think the job was completeness.
Conacher has a line for what it actually is. He says learning an accent is much more akin to learning dance steps than it is to learning a piece of music.
Explain that one, because music sounds closer.
Music you can understand completely and still not be able to perform. You can read the score, you know exactly what it should sound like in your head, and your hands won't do it. Dance is the opposite, or rather it's the same problem named honestly. You have to build it into the body. It's muscle memory. And an accent is muscle memory of a set of small movements you've never made in that combination.
So the actor who watches recordings for a month and can hear it perfectly has learned the score.
Exactly nothing. Conacher's point is that the intellectual grasp is the easy part and it buys you nothing until the mouth has been drilled.
Which is presumably where the competence stages come in.
Jill McCullough's four stages, and she's worked on Billy Elliot, The Iron Lady, No Time to Die, The Batman. Stage one, unconsciously incompetent. The actor doesn't know what he's getting wrong, which is where everyone starts, including people who are quite confident. Stage two, consciously incompetent. Now he can hear it, which feels like getting worse because every take is suddenly wrong. Stage three, consciously competent, he can produce it, but he's assembling it in real time and it costs him attention. Stage four, unconsciously competent, it just comes out.
And she says how often people reach stage four?
She says it's rare you get there. Which for a coach with that credit list is a real admission.
Working actors, and the top of the ladder is a rare condition.
There's a second thing she says that matters more for how you think about the profession. She says just because I can do it doesn't always help me teach it. You have to create a teaching system, otherwise it's just a girl in a room showing off.
That's the actual job description.
The ability and the pedagogy are separate skills and the second one is rarer. A coach is the only person on a production who ever tells an actor he's wrong. Everyone else is managing him.
So how many of them come out of linguistics? Because Daniel's guess was that they'd be plugged into the people who study speech.
The guess is partly right and the part he's missing is the identity of the specialists.
Go on.
There's no formal certification body for this. You don't get a license to be a dialect coach. The pipeline runs through drama schools, RADA, Central, and through voice studies masters programs, and what those programs have in common is phonetics. Conacher discovered he loved phonetics doing a voice studies MA. Michael Walsh has a degree in acting and a degree in linguistics, which is an unusual combination and also exactly what the job is.
So the trainers are phoneticians with a performance habit.
The core tool of the trade is the International Phonetic Alphabet. It's how you write down a sound precisely instead of gesturing at it. Walsh describes the moment it caught him, studying the IPA and then going and looking up different accents and writing out the sounds for fun.
For fun.
That's the tell. Here's the part Daniel was reaching for, though. The academic literature on accent imitation sits in phonetics journals, psycholinguistics journals, speech science. The people studying why imitation ability varies and the people teaching actors to do it are working in the same territory, and increasingly reading each other.
So the trainers are a profession that lives next door to the science and occasionally visits.
They're the applied wing. And there's a paradox Conacher points at that makes it clear how far the applied wing runs ahead of the intuition. The hardest accents to learn are the ones closest to the actor's own.
That's backwards.
Welsh actors find Geordie difficult. Geordie actors find Welsh difficult.
Say why, because my instinct says nearby should be easier.
It's the mental categories again. Two accents that are close share a lot of features, so the differences are small and local, and the ear files them as noise. That's not a different sound, that's them saying it funny. When the accent is far from your own, you hear it as foreign from the first syllable and you pay attention to the whole system. But the neighbour accent, you assume you already know it, and you stop listening.
The assumption does the damage.
Which is why the on-set job exists. A recording studio is a controlled environment. A film set is not. Audrey LeCrone describes herself as Daniel Kaluuya's ears on set, so he can focus on acting and she can catch him whenever he sounds a little British.
So there's a person whose entire job is to listen to one man.
Samara Bay sits in the creative hub behind the script supervisor, twelve-hour days, thumbs-up or a tiny sound note between takes. You can't do that work from the script. It has to be live, in the actor's ear, while he's holding an emotion, because the accent is one more thing he's managing in real time.
Twelve hours of listening for one bad vowel to show up.
And it's exactly what Conacher's dance-step framing predicts. You don't lose dance steps because you don't understand them. You lose them when you're thinking about something else.
It strikes me that this is also why the job is so invisible. Nobody watches a film and credits the person standing off-camera who caught the vowel.
The best work in this field is the work you never notice. If the coach has done the job, the audience simply believes the actor is from there. The failure mode is the opposite, you get pulled out of the scene and you can't say why. That's usually a coach who wasn't there, or an actor who didn't do the drilling.
So the reward for excellence is invisibility.
Which is a hard way to make a living, and probably part of why so few people do it.
Alright. That's the craft. The coaches can build the thing and they can build it reliably, mostly. But the question under Daniel's prompt is still sitting there, which is who gets to build it. Because everything you've described is the intervention.
And the science on individual differences is unusually direct about this. There's a 2024 study specifically on accent imitation, ninety-two imitators, over a thousand raters scoring them. The spread was enormous. Best imitator scored 4.99 out of seven. Worst scored 1.94.
That's not a talent gap, that's a canyon.
Same picture from earlier work. A 2013 study of a hundred and thirty-eight German-English bilinguals found imitation ability roughly normally distributed. Mean around 4.6 on a ten-point scale, spread from two and a half up to nearly eight.
Bell curve.
Not a gift you have or don't. And that distinction is the whole answer to Daniel's question, because a binary gift means the untalented can't be helped. A distribution means most people sit somewhere in the middle and can move.
Move how far, though. That's the part a bell curve doesn't answer.
McCullough's answer from the profession is that you know very quickly whether somebody can do this. A coach watches an actor for an hour or two and has a sense of where the ceiling is. But the training literature keeps finding the same predictors of that ceiling, and they're not what you'd guess.
Not intelligence.
Not intelligence and not memory. Musicality. That's the strongest recurring correlate. A 2019 study of German speakers faking a French accent found accent-faking ability correlated with singing ability around 0.4, tonal musicality around 0.39, rhythmic ability around 0.4.
And the memory result?
Zero correlation with working memory. Which sounds like it contradicts the intuition that you need memory to hold new sounds, and there's a resolution. Working memory seems to matter for learning the sound in the first place. Musical ability matters for retrieving it and storing it in a form you can reproduce.
So the singer has a place to put it.
And within musicality there's a finer split. Christiner and Reiterer found singers outperform instrumentalists at foreign-accent imitation. Their paper title is A Mozart is not a Pavarotti.
The instrument is the voice.
The people who train their voice, who already think in terms of pitch and placement and where a sound sits, have a head start on people who train a piano.
Which makes the whole thing look less like phonetics and more like singing.
It's not either. The 2024 study found three things feeding into it at once. Musical perception, vocal articulation, meaning how well you control the moving parts, and a personality trait, openness to experience.
Personality.
And a fourth, which is the one I keep thinking about. Imitation improved based on the imitator's attitude toward the talker. If you like him, or identify with him, you do it better.
So it's not a purely mechanical skill.
Bill Hader is quoted on exactly this in the study. He says when you're nervous around someone you start to talk like them, and that he'd done it since he was a kid, and that he thinks it lent itself to being able to do voices.
The boy copying the adults because he wanted in.
Attitude as an input. You lean toward the person and the mouth follows.
Now here's the thing I want to press on. Daniel asked whether it's a trainable skill or a remarkable aptitude, and you've said both about four times now. What does that actually look like at the individual level? A person with a low ceiling. Can they still get meaningfully better?
They can get better and stay below someone with a high ceiling who's had a week of training. That's what a distribution with a real spread gives you. The 2013 study found high-ability imitators had a measurably larger articulation space, correlated with ability at 0.7, which is a very strong relationship.
Articulation space, meaning?
The range of positions their mouth and tongue can reach. Larger space, more sounds available. Low-ability imitators showed a different signature in the brain, more diffuse activation, cortical effort, like more of the machine was being recruited to do the same job.
More people pushing the cart.
And that's a description of what practice is for. It's reducing the effort. So it's a bit of both, but not a vague bit. The base you start from is real and measurable, and it isn't a wall, it's a slope.
Herman.
Mm.
I want a real number on the motivation question, because you've said attitude matters and you've given me a personality trait and a Hader quote. Is there anything showing that the liking is doing measurable work, or is that just the study author being nice?
The 2024 study is the one that put the attitude variable in directly, and the effect was real enough to report as a finding. My honest read is that the mechanism is identification. You imitate the people you're becoming, and if you're not becoming them, the accent feels like a costume.
A costume you're wearing in front of them.
Which is why so much accent coaching is really bonding work. The coach gets the actor to like the character's world a little before anyone touches a vowel.
So the first job isn't the mouth, it's the heart.
In a manner of speaking, yes. You have to want to sound like them before you can.
Let's take one call, we've got a bit of listener mail sitting here. This is Opal.
Opal: Hi, hi, I'm calling from the queue at the pharmacy in Tel Aviv, I have been on hold so long I have memorized the hold music, so this is my window. I love the accent episodes but I think you two are missing the big one. I have lived here fourteen years and I have an accent that is nothing, it is nowhere, my mother doesn't recognize it and my Israeli friends say I sound French for some reason, and I want to know, if an actor can learn one, can I learn like, a nice one? Just a nice one. Is there a nice one I can buy? Okay, that's my question, shalom.
She wants an accent off the shelf.
What I'd tell her is that a nowhere accent is doing its job, which is being understood by everybody. And the reason nobody can place it is that she stopped converging on any one group.
So she's got the thing she asked for.
And there's a whole industry answering her actual question. Accent modification for professionals.
The industry Daniel wondered about.
Executives, physicians, professors, people on the phones all day. Usually non-native English speakers being coached for intelligibility. And the coaching is the same technique as the actors, decompose the accent, find the salient features, drill them into the mouth.
And speech-language pathologists do it too.
They run intelligibility enhancement programs, usually in a clinical setting. There's a striking finding in that literature. A record review of a hundred and seventy-five clients from thirty-five countries found only about a quarter of them reported trouble with the English th sounds.
And actually had trouble?
Ninety-five percent substituted them. So roughly three quarters of the clients couldn't hear a feature of their own speech that was audible to everyone around them.
Which is the same story as the Welsh actor and the Geordie accent.
Same story. The ear that has to judge the output is the ear that made the error. Which is the entire argument for a coach existing, and it's the same argument whether the client is a movie star or a surgeon.
Then there's the word for it.
Michael Walsh calls accent reduction a ridiculous term, because he's just teaching them a different accent, an American one. The clinics frame it as intelligibility, not erasure.
Because the framing decides what the client thinks is happening to their identity.
The honesty of the coaches on this point is striking. They're not selling authenticity. They're selling a specific, portable, learnable thing that gets you taken seriously in a room you weren't born in.
Which is a thicker subject than it first looks. Who gets told to change and who gets told their accent is charming.
Entirely different conversations depending on who's asking. And notice that the training literature doesn't pretend to be neutral about it. It's up front that the goal is intelligibility in a specific market, not correctness.
Alright, so we've got the craft and we've got the curve. Anything left on the shelf before we hand over?
One thing. Nobody has actually run the clean experiment of professional actors against the general population on this. The 2024 study had actors in it, six professionals, ten drama students, and the paper doesn't break out whether they beat everyone else.
Which is a strange hole.
It is. My suspicion is that the professionals would win on the trained accents and the amateurs might surprise people on the cold reads, but nobody's tested it, so that's a guess and I'll flag it as one.
Noted as a guess. Before we wrap, I can't hear about a two-year-old rehearsal and a control group in the same segment and not hear about Allen Whitcombe's garden again, sorry, that's my contribution.
You didn't need to bring that up.
I know.
