Can letters have a taste?

In the mind, yes: everywhere it has been tested, people read round letters as sweet and angular ones as sour, bitter or salty. In the mouth, only a trace — and in coffee it was the angular letters, not the round ones, that changed the taste.

Ebbi · 19 October 2026

Picture the words “eat me” written twice: once in soft, round letters, once in sharp, jagged ones. One of the two is sweet. You knew which before you reached the end of this sentence.

Researchers have measured that feeling, and it holds. In 2015 a team at the University of Oxford, working with the designer Sarah Hyndman, showed 101 people in the United States twelve versions of “eat me”, six rounder and six more angular, and asked how well each matched four tastes. The round ones went to sweet. The angular ones went to sour, bitter and salty. Another 99 people liked the round ones more and found them easier to read. Round, here, means rounder than letters drawn to be sharp: plain Arial was one of the round ones.

In 2018 the same group reported the test in three countries: “eat me” in Spanish, Chinese and English, shown to people in Colombia, China and the United Kingdom. In every language and every country, the round letters read sweeter. In Chinese, the round characters were liked no more than the angular ones, and found no clearer, and they still read sweeter.

So in the mind, letters have a taste. Whether they reach the tongue is another question, and far fewer tests have asked it with food in the mouth.

The jelly beans.

The first was an evening at the Victoria and Albert Museum in London, during the London Design Festival, on 14 September 2014. Eighty people were each given two identical jelly beans. They ate one while looking at “eat me” in round letters on a screen, and the other while looking at it in jagged ones. On a sheet from 0 to 10, the bean eaten with the round letters scored 7.04 for sweetness. The same bean with the jagged letters scored 5.38.

Five months later a design magazine told that experiment as a number: the people looking at the round letters, it reported, found the jelly bean 17 per cent sweeter on average. That number still travels. The only published numbers from that evening are 7.04 and 5.38. The gap between them, 1.66 points, is 17 per cent of the sheet. Not every sheet was filled in, and the researchers wrote that their results “should be treated as primarily descriptive”.

Two more evenings followed, written up with the first in 2018. At the Science Museum’s Lates, on 30 September 2015, people tasted two identical lemon-and-lime jelly beans, one beside each letter of a pair, three pairs over the evening. At the Stoke Newington Literary Festival, from 2 to 4 June 2017, 188 people tasted sour cherry and sour apple beans, and this time the difference was put to a statistical test.

Six pairs of letters, and six times the bean tasted beside the round letters was rated sweeter. Six times, too, the bean beside the angular letters was rated more sour. The widest gap is the first evening’s; the others run from a fifth of a point to just over one. At the festival, the test said the difference was real, and it was small: two thirds of a point with the thick letters, a quarter of a point with the thin ones. The beans were liked no more with either.

The researchers name the catch themselves: their tasters were “staring at the phrase ‘eat me’ written in what were obviously very different typefaces”, a set-up they call “slightly contrived”. And “it is as yet unclear”, they add, “how long-lasting are the effects”.

The coffee.

The test nearest to a shelf is about coffee. In 2020, researchers showed 146 drinkers of specialty coffee package labels that differed only in how round or angular their letters were, and then the drinkers tasted the coffee. The angular letters raised the acidity they expected, the acidity they tasted, and how likely they said they were to buy it. The round letters made the coffee taste no sweeter: “contrary to the frequently suggested correspondence between sweetness and roundness”, the researchers write, “no association was found”.

Their own reading: “different coffee profiles could be matched with different label typefaces”.

Where the letters stopped.

Two larger tests found the edge. In 2022, 480 people in Brazil picked between wrappers for a milk chocolate, sixteen in all, made of every mix of two colours, two kinds of letters, two pictures and two shapes of bar. They liked the round letters more. But the letters did not make the chocolate look sweeter or creamier. Pink made it look sweeter, and a curved bar sweeter and creamier.

The same year, a study of 220 people found the letters had “no impact on their expectations or perceptions” of sweet and sour. Among the older people, though, angular letters made the sourer food the likelier choice.

In the mind, letters have a taste. In the mouth, a trace of one.

On the label.

So don’t count on letters to sweeten your drink. In tests built to make people look at the letters, round ones lifted a jelly bean’s sweetness by a fifth of a point to a point and two thirds, out of ten.

What letters do, wherever it has been tested, is promise. Before the can is open, round letters say sweet, and angular ones say sour, bitter or salty. A sour beer in soft, round letters promises the wrong taste. So does a sweet soda in jagged ones.

And when the promise is the taste a drinker wants, it carries: the angular letters on the coffee promised acidity, the cup seemed to deliver it, and the drinkers said they were likelier to buy it.

If someone offers you letters that let you cut the sugar and keep the sweetness, ask to see the test on your own product, on a shelf, after the first week. I found none.

Seen while makingMy own letters