High sound, small ball
If someone showed you two balls, one small and one big, played a high-pitched sound and asked 'Which one is making this sound?', what would you say? Most people point to the small ball. When a lower sound plays, they point to the big one. Matches like this, between features belonging to different senses and largely shared across people, are called crossmodal correspondences.
Charles Spence's extensive review, published in 2011, is one of the field's key references. Spence lists the matches between the pitch of a sound and visual elevation, size and brightness among the most robust examples. These matches do not show up only in questionnaire answers; responses to congruent pairs can be faster than to incongruent ones.
Do toddlers also say 'high sound, small ball'?
In 2004 Catherine Mondloch and Daphne Maurer showed 30–36-month-old children two bouncing balls and asked which ball a centrally played sound belonged to. The children reliably matched the higher-pitched sound to the smaller, lighter ball. The researchers took this as a hint that some correspondences may be traces of connections present before learning.
But not every correspondence is that early or that universal. Some are learned by observing the world: big animals, big drums and big objects usually sound lower. Others draw on concepts carried by language.
Why is pitch 'thin' and 'thick' in Turkish?
In Turkish a high-pitched sound is called 'thin' (ince) and a low-pitched one 'thick' (kalın). In languages such as Dutch and English, pitch is 'high' and 'low'. In 2013 Sarah Dolscheid and colleagues compared Dutch and Farsi speakers; Farsi, like Turkish, uses thickness words. They found that the two groups behaved differently on simple, non-linguistic pitch tasks, and that once Dutch speakers had been trained to describe pitch with thickness words, they came to resemble Farsi speakers.
In 2020 Dolscheid, Çelik, Erkan, Küntay and Majid tested Turkish and Dutch speakers. Both groups matched high pitch with a thin line, but they differed on height. When a thick line was high and a thin line low, Turkish speakers mostly chose by thickness, while Dutch speakers more often chose by height. That makes the thin–thick question in this experiment especially interesting; in the crowd results we show players on the Turkish and English pages separately.
Harsh timbre, spiky shape
In the famous bouba–kiki experiment (the 'Bouba or kiki?' experiment in this lab), people match made-up words to shapes. Here there are no words: two sounds play at the same pitch, one with a soft, rounded timbre and the other buzzy and harsh. In a 2014 study with 119 people, Mohammad Adeli and colleagues found that soft timbres were strongly matched with rounded shapes and harsh timbres with angular ones.
The same study contains an interesting detail: the fundamental frequency of the sounds, that is, their pitch, was not related to the vertical position participants chose. The match between pitch and elevation has been found to be strong in many studies, but it does not appear with every stimulus and task.
The shape of sweet, the color of sour
Crossmodal correspondences also involve tastes. In the 2016 study by Carlos Velasco and colleagues, participants' judgments of taste words and shapes fell into two clusters: round shapes and 'sweet' on one side; angular shapes with 'salty', 'sour' and 'bitter' on the other. The round cluster was rated more positive, the angular cluster more 'potent'.
Colors follow a similar pattern. According to the 2015 review by Spence and colleagues, sweet goes most often with red and pink, sour with green and yellow, salty with white and blue, and bitter with black and purple. Woods and Spence largely replicated this in 2016: sweet was matched most often with pink, sour with green, salty with white and bitter with black. In our colour questions we name each taste without examples; saying 'like lemon' would suggest yellow, and 'like coffee' brown. Packaging designers are interested in these matches too: the color and shape of a package can shape what we expect its contents to taste like.
Does everyone make the same matches?
No. The field increasingly takes cultural differences seriously. In 2013 Andrew Bremner and colleagues worked with the Himba community in northern Namibia. Himba participants made the bouba–kiki match like Westerners, but they did not match carbonated water to an angular shape as Westerners do, and they tended to match less bitter milk chocolate to angular rather than rounded shapes, the opposite of Westerners.
So the 'typical direction' on the result screen is not a right answer, just the direction most often reported in research. Some of your matches may differ from the crowd; that does not mean your senses are working wrongly, only that they carry traces of your experience, your language and your preferences.