Font Weight:

Light weights

• Convey elegance, gentleness, and a modern delicate touch, but if too light can appear weak or hard to read.

Bold weights

• Signal strength, power, and confidence, but extreme heaviness can feel aggressive or reduce legibility.

Regular weight

• Best for readability, sustaining reader comfort over long periods.
• Use bold or light sparingly for emphasis, aligning the choice with the brand’s voice (e.g., bold for emphasised tone, light for a refined tone).

Addition: All of the above is aesthetic. Weight also does three documented commercial things, and none of them were in the original post.

Weight is read as physical weight. A delicate typeface symbolically carries light and thin. A bold typeface carries heavy and fat. The effect appeared specifically in consumers with a pronounced health-promotion focus, meaning it works hardest on the people actively shopping for a clean product.

Weight is read as price. Visually heavier packaging raises price expectation.

So light type is doing two jobs at once: signalling clean, and signalling cheap. If you're premium and lightweight, that's a real tension in your system, and it needs resolving somewhere other than the typeface.

Weight is the only lever for prominence. In Henderson, Giese and Cote's model, prominence requires high weight together with low naturalness, low flourish and low harmony. Weight is the single dimension that buys it, and it costs three others. Covered in full in The Typography Trade-Off

‍Width (Condensed vs. Extended):

Condensed (narrow) fonts

• Feel tight, intense, and precise – they can add energy, urgency, or a sense of efficiency.
• They have been linked to emotions like joy, fascination, and excitement in studies, but can also add an aspect of anxiety.

Extended (wide) fonts

• Evoke spaciousness, relaxation, and stability – lending an open, calm presence.
• Very wide forms can also appear slow or solemn, correlating with emotions like sadness or boredom if overused.

In branding & psychology

• ‍Condensed fonts suit modern, bold brands or tight spaces
• ‍Extended fonts signal luxury and confidence.

Correction: That last pair conflates two different things, and the second half is the wrong variable.

Width is the horizontal proportion of the letterforms. Spacing is the gap between them. The "luxury" signal people are describing when they say extended is far more reliably about generous letter-spacing, which you can apply to a condensed face and lose entirely on an extended one. The original line would lead someone to pick the wrong lever.

What actually holds up: the condensed-equals-intensity finding is corroborated from an unexpected direction. In Henderson's model, the engaging impression requires high compression alongside high elaborateness and low harmony, and that engaging factor loads positively on interesting, emotional and innovative. So compression genuinely does buy energy. It also loads negatively on calm, formal, honest and familiar, which is the part nobody mentions.

Condensed type isn't just modern and bold. It's modern, bold, and quietly costing you perceived honesty.

‍X-Height (Tall vs. Short lowercase):

High x-height

• Greatly improves legibility at small sizes.
• Projects a friendly, contemporary vibe, as more of the letter area is filled, making text blocks look solid and approachable.
• Many modern fonts use large x-heights to appear open and user-friendly.

Correction: "Greatly" was doing unearned work, and the finding is more conditional than I made it sound.

McCarthy and Mothersbaugh (2002) found that increasing x-height improves legibility for smaller typefaces only for faster readers. The same conditional applied to serifs in their data. There's also a joint effect between style and x-height that is stronger on faster readers.

That's a much more interesting finding than the one I published. X-height is not a universal legibility upgrade. It's an upgrade for people already moving quickly, which in practice means the confident, skim-reading, already-engaged part of your audience. The person struggling with your ingredient list is not who it helps.

Poulton (1972) gives the actual floor: x-height should not fall below about 6.6pt for Univers or Times New Roman, and Perpetua needs roughly 8.5pt minimum body size. Those are numbers you can hold a designer to.

Low x-height

• Lends an elegant, classic look.
• It introduces more white space within and between letters, often perceived as sophisticated or highbrow (common in luxury or traditional contexts).
• Can hurt legibility if text is small, so it’s best reserved for larger titles or logos.

The ideal choice depends on context

• For on-screen or small-print text, a generous x-height is more trustworthy simply because it’s easier to read (readers feel comfortable, thus more confident in the content).
• For a logo or headline where readability is ensured by size, a lower x-height can set a prestigious or delicate tone.
• Research confirms that a balanced, moderately high x-height contributes to a perception of a font as “legible” and even “formal” – implying professionalism with clarity.

Correction: "Research confirms" with no named research is exactly the sort of thing I've spent this year removing from my own site.

But the trustworthiness point above was right, and there's a name for it. It's called processing fluency, and it's better evidenced than almost anything else in this post. Reber and Schwarz found statements presented at high visual contrast were judged true more often than the identical statements at low contrast. Song and Schwarz found that instructions in a hard-to-read font led people to estimate a task would take longer and require more skill, and made them less willing to attempt it.

Ease of reading doesn't stay attached to the text. It transfers to the claim, and to the task the claim describes.

‍Curve Style (Rounded vs. Angular):

Rounded forms (soft curves, circular shapes)

• Evoke warmth, friendliness, and approachability.
• They subconsciously signal safety and comfort, aligning with our preference for curves over sharpness.
• Ideal for brands wanting to appear caring, playful, or inclusive (e.g., many social media and kids’ brands use rounded type to feel welcoming).

Angular forms (sharp corners, straight lines)

• Convey strength, decisiveness, and formality.
• They can feel efficient and modern, but also a bit aggressive or cold if overused.
• Suited for brands aiming for a techy, edgy, or highly professional image. For example, a cybersecurity firm might use angular letterforms to communicate authority.

‍

Major emotional differences emerge when comparing overtly curved vs. very angular typefaces in general, Small details like slightly rounded terminals vs. square terminals showed no significant emotional difference.

This means it’s the overall impression (how curvy or not the font looks at a glance) that matters more than minor design flourishes.

Fonts are not just aesthetic tools; they carry emotional weight and can reinforce (or undermine) a brand’s message in the mind of the consumer. The best results come from aligning typeface characteristics with the intended brand personality and values – and backing those decisions with evidence on how viewers actually respond.

Correction, and this is the significant one: terminals matter. The original claim has been overtaken.

Velasco et al. (2018) found that when bold strokes are combined with angularity, there's an enhancing effect on intense tastes such as bitterness, and concluded that designers should configure curvature and weight together rather than treating them as separate decisions. That's an interaction effect, which is precisely what the struck paragraph denied.

A 2026 study in Behavioral Sciences went further, testing four Chinese typeface categories against three stroke thicknesses and finding a significant interaction between typeface category and stroke thickness across all five flavour dimensions. The mechanism they identify is partly the bifeng, the structure at stroke terminals. Terminals, specifically.

So the correct version is: overall curvature sets the direction, and terminals plus weight modulate the intensity. They are not independent, and the small details are not noise.

‍

New for 2026: Your font is telling people what your serum feels like

This is the section the original post should have had.

The finding

People systematically match rounder typefaces with sweet, and angular typefaces with bitter, salty and sour. That's been replicated repeatedly since Velasco et al. (2014), and it's held across writing systems.

A three-language, three-country study tested Spanish, English and Chinese typefaces with participants in Colombia, the UK and China. Rounder typefaces were evaluated as sweeter than angular ones in every language and every country. The angular-equals-bitter-salty-sour effect appeared in Spanish and English but did not replicate for Chinese typefaces, which is a useful reminder that the mapping travels less reliably than the direction does.

The 2026 Chinese study found the rounded Yuan typeface with thin strokes tended toward sweetness, bold Song typefaces tended toward bitterness or spiciness, and bold Hei typefaces tended toward saltiness. Curvature-sweetness and angularity-non-sweetness held across both writing systems.

The finding that should stop you

Expectation is one thing. This is another.

In a specialty coffee study, an angular typeface on the packaging label increased both the expectation and the actual perception of the coffee's acidity, and increased purchase intent.

The typeface changed what people reported tasting. Not what they guessed they would taste. What they said they tasted, after drinking it.

The axis underneath it

Analysis of these correspondences produces a semantic space with two principal components, which the researchers describe as roughly potency and evaluation, and two clusters. One cluster contains round shapes and the taste word sweet. The other contains angular shapes and the words salty, sour and bitter. The round cluster is more positively valenced and less potent. The angular cluster is more potent and less positively valenced.

Read that as a brand strategist rather than a psychologist and it says something specific.

Round buys likeability and costs perceived potency. Angular buys perceived potency and costs likeability.

Why this is the most useful thing on this page if you sell skincare

Taste is the wrong sense for our category. Potency is not.

Potency is the entire proposition of an actives-led product. It's what a retinol, an acid, a vitamin C is selling. And the research says angularity is how visual design communicates potency, at the cost of warmth.

Which sets up the real decision:

Barrier repair, sensitive skin, soothing, ceramides. You need the round cluster. Warm, gentle, low-potency signalling, because a customer with a compromised barrier is buying reassurance and reads potency as risk. Rounded forms, lighter weight, soft terminals.

Actives, resurfacing, clinical, results-led. You need the angular cluster. The customer is buying efficacy and reads gentleness as this won't work. Angular forms, heavier weight, sharp terminals. You will pay for it in warmth, and you should spend that budget knowingly.

The failure mode: a potent formula wrapped in round, light, gentle type, and a founder who can't work out why the reviews say it's lovely but nobody believes it does anything. The packaging is arguing with the formula.

And there's supporting evidence that this extends past taste into texture. A study of 3D-printed surfaces found rounded textures were linked with bouba and with pleasantness, while pointed textures were associated with kiki, higher arousal and warmer colours. Separate work finds high correspondence between visual and tactile evaluations of material roughness and hardness. The visual-to-tactile channel is real.

The caveat, because this is an inference

The direct evidence is on taste, and on visual-to-tactile judgements of roughness and hardness. There is no study I've found that puts an angular typeface on a serum and measures whether people report a stronger tingle.

What I'm doing above is extending an established mechanism into an adjacent sensory domain where the underlying logic holds. That's a defensible hypothesis and a good basis for a design decision. It is not a finding, and I'm not going to present it as one.

If anyone runs that study, I'd like to read it.

What this post still doesn't cover

Letter-spacing and letter case. Both are genuine small details with real effects, both were missing from the original, and both are missing from this update because I haven't finished reading the research properly. There's an established effect where all-uppercase brand names are perceived as more premium than lowercase ones, mediated by conspicuousness and moderated by how socially visible the product is. When I've read the paper rather than the abstract, it'll go in.