We were wrong about the squinch
This page has been claiming that the photographer Peter Hurley’s squinch is the muscle action psychologists call the Duchenne marker. We were pleased with that, because nobody else had said it. That should have been the warning.
The coding system for facial movement separates two actions in the same muscle. Action Unit 6 is the cheek raiser, in the outer ring of the eye muscle. It lifts the cheek and makes the crow’s feet. That is the Duchenne marker, and it is what the entire research literature is about.
Action Unit 7 is the lid tightener, in the eyelid part of the same muscle. The lower lid pushes up, the effect gathers toward the inner corner, and the cheek does not move.
Now here is Hurley describing the squinch, in his own words.
“The Squinch only uses the lower eyelid. You raise it just enough to bring it slightly closer to your pupil. The upper eyelid barely moves.”
“Anatomically, you’re narrowing the palpebral fissure… predominantly from the bottom rather than evenly from both sides.”
No cheek. No crow’s feet. The palpebral fissure is the eyelid opening. That is a description of Action Unit 7, not Action Unit 6.
Why this matters more than a labelling error. The whole point of the claim was to borrow the Duchenne research and apply it to the squinch — to say that a photographer’s trick is really a documented muscle action with measured effects on how people are judged. If the squinch is a different action, none of that evidence transfers. The studies are about the cheek raiser. So we are left with a technique that many photographers swear by, taught by somebody who has shot a great many faces, with no research behind it at all. That may still be worth doing. It is not what we said it was. We had also convinced ourselves that the absence of anyone else making this connection was evidence we had spotted something. It was evidence of the opposite.
There is a second problem we should have noticed. The Duchenne marker is prized precisely because it is hard to produce on purpose. A technique you can drill in a mirror is a strange candidate for the one muscle action the literature values for resisting deliberate control.
Can you produce a genuine-looking smile on purpose?
The research here is real and it is more interesting than the version we were given, because the striking number turns out to come from a much easier task.
Two older studies found that almost everybody could do it. Told simply to appear happy, one hundred per cent of participants produced the marker in a 1996 study. Told to pose joy, ninety-one per cent did in a 2006 one. On that evidence, the famous claim that the Duchenne smile cannot be faked is wrong.
But the seventy-one per cent figure we were given does not belong with those, and putting it in the same range would have been misleading. It comes from a 2013 study — our research dated it to 2015 — which ran two different tasks. When participants were simply asked to smile as though genuinely happy, only about a third produced the marker. Asked to fake positive emotion, about a quarter did. The seventy-one per cent comes from a separate condition where they were shown a photograph of a Duchenne smile and asked to copy it. Copying a picture in front of you and generating an expression from a word are not the same skill, and quoting them as one range from seventy-one to one hundred per cent would have hidden that completely.
What survives is still useful. Most people can produce the marker when told to look happy, and rather fewer can produce it reliably from a standing start. It is trainable, it varies a lot between individuals, and copying a reference image is the easiest route in.
And one claim we are dropping entirely. Our research offered a striking piece of history: that Paul Ekman’s own group found most people could produce these expressions on request, called the result “quite unusual”, and never mentioned it again. We could not reach the source, which is an unpublished dissertation, and everything we could check points the other way — Ekman’s published position is that the ability is rare, and the surrounding literature describes his group screening for the small minority who have it. That is a serious allegation about a named scientist’s conduct and we are not making it on a paraphrase of a source we cannot open. If the dissertation says what we were told, it deserves a proper reading first.
Does the marker actually do anything?
Yes, and there is a meta-analysis. There is also a caveat inside it that matters enormously for a headshot specifically.
Gunnery and Ruben pooled the research in 2016 and the finding is quotable exactly as they wrote it. Duchenne smiles and the people producing them are “rated more positively (i.e., authentic, genuine, real, attractive, trustworthy)” than non-Duchenne smiles and the people producing them.
And the effect survives being deliberate. In a 2020 study the authors concluded that although the marker can be posed without positive feeling, “it is still perceived by others to be indicative of genuine emotion”. So the question of whether anyone can tell is the wrong question. The marker does its work either way.
The caveat our research did not mention, and it points straight at your headshot. The meta-analysis found the effect was moderated. It was larger for video than for photographs. It was larger for naturally occurring smiles than for posed ones. And it was larger when the two kinds of smile were not matched for how hard the mouth was working. A headshot is a posed still photograph. That is the combination of conditions in which the measured advantage is at its smallest. The effect is real and we are not going to quote it at you as though a photograph and a video were the same thing.
It is not read the same way everywhere
This closes a gap our research flagged and could not fill, and the answer is more definite than expected.
A 2012 study by Hess and colleagues carries its finding in the title: the Duchenne marker is not a universal signal of smile authenticity, but it can be learned. Canadian participants both produced and relied on it. Gabonese participants did not use it to judge authenticity at all. Chinese immigrants to Canada showed sensitivity to it when judging French-Canadian faces, which suggests the reliance is picked up rather than built in.
And a genuine absence, which we would rather report than fill with reasoning. We went looking for research on whether the marker reads equally across ages and skin tones. On skin tone we found nothing at all — no study of how human viewers perceive this marker on darker skin, where lower contrast may make fine wrinkling less visible. The only adjacent work is about facial recognition software performing worse on darker faces, which is a different question about machines. On age, there is nothing direct either, though related work suggests smile wrinkling varies with age and sex. Given that the marker is defined by visible wrinkling, this seems like an obvious thing for somebody to have studied. Nobody has.
The Botox study, properly reported
Our research promised this study and did not say what it found. It found something clear, and two things about it need saying alongside.
Photographs of thirty-two patients were rated before and about a month after an injection that paralyses the muscle at the outer corner of the eye. With the muscle working, their smiles were rated more felt, more spontaneous, more intense and happier. The differences were small but consistent, and statistically clear.
What we were toldWhat the paper saysThe correction
Effects on attractiveness, dominance and intelligence were judgment-dependent
The words dominance and intelligence do not appear anywhere in the paper. It measured attractiveness, which showed no significant change, and perceived age.
That finding belongs to a different 2013 study. Two papers had been spliced together in our source.
A neutral scientific test
It was funded by the company that makes the product, as a late-stage clinical trial, using patients from that company’s own study.
That does not make the finding wrong. It is the kind of thing a reader is entitled to know.
Evidence about genuine smiles
The authors note that only deliberately posed smiles were photographed, in still images rather than video, with a US-only rating sample.
It cannot speak to spontaneous smiles at all, and the authors say so themselves.
We are reporting it flatly and without moralising. It is a fact about one muscle and one procedure, and the muscle happens to be the one this chapter is about. If your face is your instrument, that is worth knowing before rather than after.
What actually makes eyes look alive
With the squinch claim gone, this section carries the page. It survives, because it never depended on the research in the first place.
There is one piece of evidence worth having, and we tracked it to its source rather than trusting the summary we were handed. In a 2001 eye-tracking study of fifty-eight people, viewers made significantly more fixations on the outer eye region for happy faces than for neutral ones. People look at the eyes. That is a small checkable fact underneath a large amount of assertion.
Direct the cause, not the effect
You cannot arrange your face into looking interested. You can decide what you are interested in. The face follows, and this is the same principle as every acting chapter on this site.
Put a person behind the lens
The acting teacher Mae Ross, writing in Backstage: “Act as if the camera is your best friend, your true love, or your enemy” — depending on the expression you want. A lens is not a person and your face knows it.
Let the light do its half
A catchlight is the reflection of the light source in the eye. Without one, an eye at thumbnail size is a dark patch. No amount of thinking replaces the physics.
The test that decides itPhotograph yourself thinking about a specific person, then photograph yourself trying to look warm. Show both to somebody who knows you and ask which one you were thinking in. They will get it right.
On nerves, one useful cross-reference from elsewhere in the Academy. Frightened stillness and attentive stillness are physiologically different things, and the difference shows. Press your feet into the floor, breathe out longer than you breathe in, look at one real object, and then look at the person. The last twenty minutes of a session are the good ones for a reason.
The geometry, and how little of it is real
People arrive at this page wanting chin angles. Here is everything we could actually stand up, and everything we could not.
Forehead forward, not chin outPush the head forward and very slightly down. It stretches the line under the chin and defines the jaw. It takes four seconds, it is free, and photographers describe the same mechanism independently of each other.
Real
Hands, even out of frameMae Ross is specific: relax them, never make fists, and if they are above the waist let the fingers point up and the wrists break. Nothing you do with your hands will be seen. Everything you do with them will be felt.
Real
Tilt your headNo portrait authority and no casting professional was found saying this. There is one clinical study about tilting the head downward, from surgical photography, which is a different thing from the sideways tilt people mean.
Folklore
Never square-on, always three-quartersNo source. And the one relevant quote goes the other way — a headshot photographer writing in a trade publication says a three-quarter shot will not stand out as well in an online gallery.
Folklore
The correct camera heightGeneral portrait advice converges loosely on eye level or a little above, all of it from unbylined how-to pages. No named authority states a rule for actor headshots. We said last time that we did not know, and we still do not.
Unknown
Teeth or no teethWe have now looked twice. No named casting director is on record either way, which is itself the answer. Anyone telling you there is a rule is telling you their preference.
Unknown
One thing worth adding from a talent manager, Terrie Snell, who was asked what she looks for. Whether there is a clearly defined emotion reading in the shot, and whether the eyes are alive. Not an angle. Not a crop. The same answer everybody in casting gives when you actually ask them.
Grading the sources
This is the first page in this series where the thing we had to correct was our own original idea rather than somebody else’s research.
The claim, and where it came fromWhat we found
Duchenne smiles are rated more authentic, genuine, real, attractive and trustworthyGunnery and Ruben meta-analysis, 2016
Primary
Quotable word for word from the abstract. Moderated by video versus photograph and natural versus posed, which our research did not mention and which matters for a headshot.
A posed marker is still read as genuine emotionFrontiers in Psychology, 2020
Primary
Verbatim from the discussion. Same paper as the Botox study, which our research had cited as two separate sources.
The marker is not universal and can be learnedHess and colleagues, 2012
Primary
Peer-reviewed and directly on point. Closes a gap our research flagged and could not fill, and it says the reliance is cultural.
Viewers fixate on the outer eye region for happy facesAn eye-tracking study, 2001
Primary
Traced to the real paper. Our research had it from an undergraduate class essay uploaded to a document-sharing site, which happened to summarise it correctly.
Voluntary and spontaneous smiles use different neural routesStandard neuroanatomy
Good
Well established, resting on patients who lose one and keep the other. A simplification, since the two routes interact, but a fair one.
What to do with your handsMae Ross, Backstage
Good
Named, dated, specific and on the record. The only real source we found on a question everybody asks.
You are an actor, not a modelA photographer’s own site
Check
The sentiment is hers and it is all over her material. That exact wording we could not find, so we have not put it in quotation marks.
Actors should face left in the frameAn agent, on a photographer’s marketing blog
Check
A real quote from a real person, on a page with no confirmable original date. Second time we have flagged this source. Interesting, not load-bearing.
The squinch is Action Unit 6, the Duchenne markerOur own original claim
No
It is a lower-lid action with no cheek involvement, which is Action Unit 7. Hurley’s own description says so. None of the Duchenne research transfers.
Between 71 and 100 per cent of people can produce the marker deliberatelyThree studies, presented as a range
No
Not one range. The 71 per cent came from copying a photograph. The same study’s from-instruction figure was under a third, and it was misdated by two years.
Ekman’s group buried a contradicting resultAn unpublished dissertation, at second hand
No
We could not open the source and the surrounding evidence points the opposite way. Not publishing an allegation about a named scientist on a paraphrase.
The Botox study found judgment-dependent effects on dominance and intelligenceOur research
No
Those words appear nowhere in the paper. That finding is from a different 2013 study, spliced in somewhere upstream of us.
One thing to say plainly about the correction at the top. Nothing in the Duchenne literature is wrong, and nothing about it stops being interesting. What changed is that it is research about smiling rather than research about a photographer’s technique. Those were two subjects and we had glued them together.
Where we differ from the standard account
Is the squinch the Duchenne marker?
Commonly saidWe said so ourselves, and were pleased about it.
What we foundNo. It is a lower-lid action, and the Duchenne marker is a cheek action. Different codes, different muscle regions.
Can a genuine smile be faked?
Commonly saidFamously not. That is the whole appeal of the idea.
What we foundMost people produce the marker when told to look happy. Rather fewer can do it reliably without a picture to copy.
Does it still work if it is deliberate?
Commonly saidSurely not, if it can be faked.
What we foundYes. A posed marker is still read as genuine emotion. But the advantage is smallest in posed still photographs.
Is the marker universal?
Commonly saidTreated as a fact about human faces everywhere.
What we foundOne group of participants did not use it to judge authenticity at all. The reliance appears to be learned.
Is there a right head angle?
Commonly saidEveryone has a number and a rule.
What we foundNo credible source states one. The strongest quote we found argues against the three-quarter turn.
Teeth or no teeth?
Commonly saidConfidently, one way or the other.
What we foundNo named casting director is on record either way. We have now checked twice.
What should you actually do?
Commonly saidRelax, be yourself, and smile with your eyes.
What we foundThink toward one specific person. The face is the effect, and you can only direct the cause.
Try this in front of a phone tonight. Take one photograph while deliberately arranging your face into something warm, then take another while genuinely thinking about somebody you are glad to see. Do not look at them yourself. Send both to a friend and ask which one you were thinking in. That answer is the entire subject of this page.
