Spend any time in online appearance discussions and you'll encounter a technical vocabulary: canthal tilt, midface ratio, gonial angle, ogee curve, hunter eyes, maxillary projection. The terms sound clinical, which lends them authority.
Some are genuine anatomical measurements borrowed from facial surgery and orthodontics. Others are community coinages with no clinical standing. Almost all are discussed with far more confidence about their importance than the evidence supports.
This guide defines each term accurately, explains what it actually measures, and states plainly what the research does and doesn't establish. Consider it a translation layer — so you can read the conversation critically rather than absorbing its assumptions.
First, the framework that should govern all of this
Before the individual features, one piece of context determines how much weight any of them deserves.
The best-established finding in facial attractiveness research is not about individual features at all. In Psychological Science, Judith Langlois and Lori Roggman demonstrated that mathematically-averaged composite faces are rated more attractive than almost all of the individual faces used to construct them — and that attractiveness increases as more faces are averaged in.
Gillian Rhodes's review in the Annual Review of Psychology confirmed through critical review and meta-analyses that averageness, symmetry, and sexual dimorphism are attractive in both male and female faces and across cultures.
Notice what's absent from that list: any individual feature. The research identifies whole-face structural properties, not a checklist of optimal parts. Faces are processed holistically — the visual system encodes configuration and relationships, not a sum of independently-evaluated components.
This matters because feature-focused discourse inverts the finding. It takes a literature about global properties and reconstructs it as a parts inventory. A face with an "ideal" canthal tilt and an "ideal" gonial angle that don't cohere is not attractive; harmony between features consistently outperforms optimisation of any single one.
Hold that in mind for everything below.
The eye region
Canthal tilt
What it is: the angle formed between the inner corner of the eye (the medial canthus) and the outer corner (the lateral canthus), measured against horizontal.
- Positive canthal tilt — outer corner sits higher than inner corner.
- Neutral — corners approximately level.
- Negative canthal tilt — outer corner sits lower than inner corner.
What's real: canthal tilt is a legitimate anatomical measurement used in oculoplastic and facial plastic surgery. It's assessed clinically before eyelid procedures, and it does change with age — the lateral canthus tends to descend over decades as periorbital tissues lose support. In that clinical sense, a positive tilt is associated with a more youthful periorbital appearance.
What's overstated: the online treatment of canthal tilt as a dominant attractiveness variable substantially outruns the evidence. It isn't among the traits identified in the major meta-analytic reviews. The population range is narrow — most people fall within a few degrees of neutral — and differences that small are not reliably perceptible in ordinary viewing.
There's also a serious measurement problem. Canthal tilt is highly sensitive to head position: tilting your head a few degrees, or shooting from slightly above or below eye level, changes the measured angle. A large share of people who conclude they have "negative canthal tilt" from a phone photo are measuring their head angle.
"Hunter eyes"
What it is: a community term, not an anatomical one. It describes a compound appearance — deep-set eyes, prominent brow ridge, positive canthal tilt, and relatively low upper-eyelid exposure.
What's real: the underlying features exist. Orbital depth and brow projection are genuine skeletal variations, and they're partly dimorphic — more pronounced brow projection is typically male-associated, which connects it to the sexual dimorphism literature.
What's overstated: the term itself has no clinical standing and no research base. It bundles several distinct traits into a single ideal and treats it as a universal target. The dimorphism research supports dimorphic cues mattering; it does not support one specific compound eye configuration as optimal.
Eye spacing
What it is: interocular distance — the horizontal separation between the eyes — usually expressed relative to face width.
What's real: this one has direct experimental support, and it's the best-evidenced item in this article. Pallett, Link and Lee's study in Vision Research tested facial feature arrangement systematically across four experiments and found attractiveness optimised when the horizontal distance between the eyes was approximately 46% of face width (alongside a vertical eye-to-mouth distance of about 36% of face length).
The important caveat: the researchers found that these optimal values correspond to the proportions of an average face. This isn't an exotic ideal — it's the population centre. As Pallett noted, the search for a special ratio had failed since antiquity, and what the data showed instead was that average distances form the true optimal ratios.
You can measure your own against these values with our golden ratio face test.
The midface
Midface ratio
What it is: the vertical distance from the pupil line to the base of the nose, expressed relative to face width or to overall face length. It captures whether the middle third of the face is proportionally long or short.
What's real: vertical facial proportion is a genuine variable in facial anthropometry, used in orthodontic and surgical planning. Pallett and colleagues' eye-to-mouth finding (36% of face length) is a directly relevant experimental result, since midface length is a major determinant of that distance.
What's overstated: the online treatment of "midface" as a make-or-break trait. It's one proportional dimension among several, and the optimum — as with eye spacing — is near the population average rather than at an extreme. "Short midface" is often presented as an unqualified good; the evidence supports typical, not minimal.
Maxillary projection
What it is: how far forward the upper jaw (maxilla) sits, influencing cheekbone support and the position of the midface relative to the rest of the profile.
What's real: maxillary position is a genuine skeletal variable, clinically significant in orthodontics and orthognathic surgery, and it does affect facial appearance meaningfully. Deficient maxillary projection has real functional as well as aesthetic implications.
What's overstated: claims that it can be modified non-surgically. "Mewing" — sustained tongue posture against the palate — is widely promoted as a method of improving maxillary position. Skeletal change in adults from tongue posture is not supported by evidence. Whatever effects people report are more plausibly explained by posture, soft-tissue positioning, and photographic variation than by bone remodelling.
The lower face
Gonial angle
What it is: the angle at the corner of the mandible where the ascending ramus meets the horizontal body of the jaw. A more acute angle produces a sharper, more defined jaw corner; a more obtuse angle produces a softer one.
What's real: gonial angle is a standard cephalometric measurement in orthodontics, with normative ranges. Jaw structure is dimorphic, and sexual dimorphism is one of the traits Rhodes's meta-analytic review identified as attractive across cultures.
What's overstated: the precision with which people assess it from photographs. Gonial angle is genuinely measured on lateral cephalometric radiographs, not from selfies. Soft tissue — including masseter muscle bulk and subcutaneous fat — substantially obscures the underlying skeletal angle. A photograph shows you the soft-tissue silhouette, not the bone.
Jawline definition
What it is: how sharply the jaw's outline separates from the neck.
What's real: this is the most modifiable item in this article, which makes it worth more attention than the skeletal measurements. Definition depends heavily on subcutaneous fat, which responds to body composition; on posture, since forward head position compresses the submental area; and on hydration and sodium intake in the short term.
What's overstated: devices and exercises marketed for jawline definition. Chewing devices increase masseter bulk, which widens the lower face rather than sharpening the outline — often the opposite of the stated goal.
Chin projection
What it is: how far forward the chin sits relative to the rest of the facial profile.
What's real: a genuine profile variable, dimorphic, and one of the more common targets in facial aesthetic surgery.
What's overstated: as with everything else here, the assumption that more is better. The evidence supports proximity to typical proportions, not maximisation.
Ogee curve
What it is: the S-shaped contour visible in three-quarter view, running from the cheekbone through the slight hollow below it and out again over the front of the cheek.
What's real: the ogee curve is a genuine concept in facial aesthetic assessment, particularly in filler and facelift planning. It's substantially a soft-tissue property, and it flattens with age as midface volume descends.
What's overstated: its evaluability from frontal photographs. It's a three-quarter-view feature by definition, and it's highly lighting-dependent — the same face under overhead light and diffuse frontal light shows markedly different apparent contour.
What actually holds up
Consolidating the honest position:
Well supported. Whole-face properties: averageness, symmetry, sexual dimorphism, skin quality. These are the traits that survive meta-analysis across cultures.
Directly supported by experiment. Eye-to-mouth distance at approximately 36% of face length; interocular distance at approximately 46% of face width. Both correspond to average proportions.
Real anatomy, uncertain attractiveness weight. Canthal tilt, gonial angle, maxillary projection, midface ratio. These are legitimate clinical measurements. Their independent contribution to attractiveness ratings is not established at anything like the confidence with which they're discussed online.
Genuinely modifiable. Jawline definition (via body composition and posture), skin quality, and — decisively — photographic conditions. See how to take the best photo; camera distance alone changes measured facial proportions substantially.
Not supported. Skeletal change in adults from tongue posture. Any single feature as the determining variable. Precision assessment of skeletal angles from selfies.
The measurement problem nobody mentions
One issue undermines almost all casual feature assessment: you cannot reliably measure these from a phone photo.
Head rotation of a few degrees changes canthal tilt, apparent eye spacing, and jaw symmetry. Camera distance distorts proportions — a photo at roughly 12 inches makes the nasal base appear about 30% wider relative to the face than the same face at five feet, per the mathematical model published in JAMA Facial Plastic Surgery. Lighting direction changes apparent bone structure dramatically. Expression shifts landmark positions.
Clinical facial measurement uses standardised conditions — fixed distance, controlled lighting, defined head position, often radiographic imaging for skeletal values. The gap between that and a bathroom selfie is enormous, and it's why so much online feature analysis reaches confident conclusions from unreliable inputs.
If you want proportional measurements, take them from a properly-taken frontal photograph. Our attractiveness test reports the underlying values rather than a bare composite score, and what is the average attractiveness score explains how to read them.
A closing note on proportion
There's a reason this article keeps returning to averages. The most robust finding in the field — replicated since 1990 and confirmed experimentally in the proportional work — is that the optimum sits at the population centre, not at an extreme.
Feature-focused content inverts this. It presents each trait as something to maximise: sharper angle, more positive tilt, shorter midface, more projection. The evidence points the other way. What predicts attractiveness is proportional coherence — features that fit together and sit near typical values — not any individual measurement pushed to an extreme.
And beyond all of it sits the individual layer. Germine and colleagues' twin study in Current Biology found that most reliable variation in individual face preference traces to each person's unique environment and life experience, not to shared standards. No feature analysis reaches that. It's the largest single component of whether a particular person finds you attractive, and it isn't in your gonial angle.
Frequently Asked Questions
What is canthal tilt?
The angle between the inner and outer corners of the eye relative to horizontal. Positive tilt means the outer corner sits higher. It's a legitimate clinical measurement used in oculoplastic surgery, but its importance to attractiveness is substantially overstated online, and it's very sensitive to head position in photographs.
Is negative canthal tilt bad?
No. The population range is narrow, differences of a few degrees aren't reliably perceptible, and canthal tilt doesn't appear among the traits identified in major meta-analytic reviews of facial attractiveness. Many apparent negative tilts are measurement artefacts from head angle or camera position.
Does mewing actually work?
Skeletal change in adults from tongue posture is not supported by evidence. Reported changes are more plausibly explained by posture, soft-tissue positioning, body composition, and photographic variation.
How do I get a more defined jawline?
Jawline definition depends mainly on subcutaneous fat, posture, and short-term factors like hydration and sodium. Body composition and head posture are the levers that work. Chewing devices increase masseter bulk, which widens the lower face rather than sharpening its outline.
Can I measure my gonial angle from a photo?
Not accurately. Gonial angle is properly measured on lateral cephalometric radiographs. Soft tissue obscures the skeletal angle, and photographic conditions introduce substantial error.
References
- Langlois, J. H., & Roggman, L. A. (1990). Attractive Faces Are Only Average. Psychological Science, 1(2), 115–121. https://doi.org/10.1111/j.1467-9280.1990.tb00079.x
- Rhodes, G. (2006). The Evolutionary Psychology of Facial Beauty. Annual Review of Psychology, 57, 199–226. https://doi.org/10.1146/annurev.psych.57.102904.190208
- Pallett, P. M., Link, S., & Lee, K. (2010). New "golden" ratios for facial beauty. Vision Research, 50(2), 149–154. https://doi.org/10.1016/j.visres.2009.11.003
- Little, A. C., Jones, B. C., & DeBruine, L. M. (2011). Facial attractiveness: evolutionary based research. Philosophical Transactions of the Royal Society B, 366(1571), 1638–1659. https://doi.org/10.1098/rstb.2010.0404
- Germine, L., Russell, R., Bronstad, P. M., et al. (2015). Individual Aesthetic Preferences for Faces Are Shaped Mostly by Environments, Not Genes. Current Biology, 25(20), 2684–2689. https://doi.org/10.1016/j.cub.2015.08.048
- Ward, B., Ward, M., Fried, O., & Paskhover, B. (2018). Nasal Distortion in Short-Distance Photographs: The Selfie Effect. JAMA Facial Plastic Surgery, 20(4), 333–335. https://doi.org/10.1001/jamafacial.2018.0009